Changelog

107 applied updates · RSS feed

Every entry below is a change to the index, published with the sources it was checked against. Figures are cross-checked against ≥2 independent sources before they reach the leaderboard, and anything that fails those checks is held back rather than published. Entries badged auto-approved were published without a human reading them first.

Spec update Update T-HR3 high confidence auto-approved

2026-08-29T21:48:47+00:00

Field	Before		After
payload_kg	—	→	2

100.0/100 validated · 3/3 citations resolved

✓ Schema record matches the schema
✓ Source count 3 source(s) cited
✓ Source independence 3 independent domain(s): grabarobot.com, humanoidhub.ai, max-robotics.com
✓ Citations resolve 3 of 3 cited URLs resolved
✓ Citations on-topic 3 of 3 fetched page(s) mention T-HR3
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
✓ Second-model check checked independently by deepseek: all 1 claim(s) supported

3 sources: www.humanoidhub.ai www.max-robotics.com grabarobot.com View robot →

Spec update Update RHP Kaleido 9 high confidence auto-approved

2026-08-29T21:48:13+00:00

Field	Before		After
ai_model	—	→	Physical AI, 3D LiDAR/stereo SLAM, simulation-based reinforcement/imitation learning, dynamic balance recovery
onboard_compute	—	→	Embedded real-time controllers; NVIDIA Isaac Sim for simulation
total_dof	—	→	36
dexterous_hands	False	→	True
deployments		→	Kawasaki KAWARUBA hub, Future Lab HANEDA

100.0/100 validated · 4/4 citations resolved

✓ Schema record matches the schema
✓ Source count 4 source(s) cited
✓ Source independence 4 independent domain(s): biggo.com, humanoid.press, nexthumanoid.com, robothub.app
✓ Citations resolve 4 of 4 cited URLs resolved
✓ Citations on-topic 4 of 4 fetched page(s) mention RHP Kaleido 9
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
? Second-model check second model returned unparseable output

4 sources: finance.biggo.com nexthumanoid.com humanoid.press www.robothub.app View robot →

Company update Update Unitree Robotics auto-approved

2026-08-29T21:44:52+00:00

Field	Before		After
valuation_usd_billions	6.2	→	9.0

100.0/100 validated · 3/3 citations resolved

✓ Schema record matches the schema
✓ Source count 3 source(s) cited
✓ Source independence 3 independent domain(s): co.ke, cztv.com, kuwaittimes.com
✓ Citations resolve 3 of 3 cited URLs resolved
✓ Citations on-topic 2 of 3 fetched page(s) mention Unitree Robotics
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
✓ Second-model check checked independently by deepseek: all 1 claim(s) supported

3 sources: streamlinefeed.co.ke www.cztv.com kuwaittimes.com

Spec update Update NEO (Gamma) medium confidence auto-approved

2026-08-29T21:43:28+00:00

Field	Before		After
deployments	EQT (up to 10,000 units across portfolio, 2026-2030), Consumer pre-orders (10,000+ reservations, late 2026 US deliveries)	→	NEO factory in Hayward, California fully operational since April 2026, capacity 10,000 units/year, Pre-orders opened October 2025, 10,000+ units in first week

57.5/100 validated with warnings · 1/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): qq.com, xianjichina.com
! Citations resolve 1 of 2 cited URLs did not resolve
✓ Citations on-topic 1 of 1 fetched page(s) mention NEO
✓ Value plausibility all numeric values within plausible ranges
! Internal consistency payload (25 kg) is unusually high for a 30 kg robot
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: baidu-mip.xianjichina.com news.qq.com View robot →

Company update Update 1X Technologies auto-approved

2026-08-29T21:43:27+00:00

Field	Before		After
valuation_usd_billions	10	→	6

85.0/100 validated with warnings · 1/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): qq.com, xianjichina.com
! Citations resolve 1 of 2 cited URLs did not resolve
✓ Citations on-topic 1 of 1 fetched page(s) mention 1X Technologies
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
✓ Second-model check checked independently by glm: all 1 claim(s) supported

2 sources: news.qq.com baidu-mip.xianjichina.com

New model New model: Xiao Di (BYD) high confidence auto-approved

2026-08-28T04:59:31+00:00

status commercial autonomy supervised-autonomy DoF 31 payload — kg deployments 3

BYD's first humanoid robot, Xiao Di, is a 1.61m, 58.5kg service robot with 31 degrees of freedom and dexterous hands (7 DOF per hand) designed for showroom customer interaction. It debuted at Di Space experience centers in August 2026, with plans to deploy 2-3 units per dealership across 50 locations.

70.0/100 validated with warnings · 5/6 citations resolved

✓ Schema record matches the schema
✓ Source count 26 source(s) cited
✓ Source independence 18 independent domain(s): ai2.work, baidu.com, chosun.com, ebrun.com, embodiedglobal.com, ienvi.com.au, instadatanews.com, kr-asia.com, longbridge.com, nextfin.ai, roboticsintl.com, robotsbeat.com, science.report, smartbrainsai.com, sznews.com, techfastforward.com, tesorb.com, toutiao.com
! Citations resolve 1 of 6 cited URLs did not resolve
✓ Citations on-topic 4 of 5 fetched page(s) mention Xiao Di
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim 'supervised-autonomy' not contradicted by the cited evidence
! Second-model check checked independently by glm: 5 claim(s) supported, 1 not addressed by the evidence

26 sources: sztqb.sznews.com techfastforward.com techfastforward.com youjia.baidu.com View robot →

Spec update Update REEM-C high confidence auto-approved

2026-08-28T04:58:38+00:00

Field	Before		After
dexterous_hands	True	→	False

72.5/100 validated with warnings · 3/3 citations resolved

✓ Schema record matches the schema
✓ Source count 3 source(s) cited
✓ Source independence 3 independent domain(s): grabarobot.com, leobotics.com, renishaw.com.cn
✓ Citations resolve 3 of 3 cited URLs resolved
✓ Citations on-topic 3 of 3 fetched page(s) mention REEM-C
✓ Value plausibility all numeric values within plausible ranges
! Internal consistency status 'commercial' with autonomy 'research' is contradictory
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

3 sources: www.grabarobot.com en.leobotics.com www.renishaw.com.cn View robot →

New model New model: Kangaroo (PAL Robotics) medium confidence auto-approved

2026-08-28T04:58:38+00:00

status prototype autonomy research DoF — payload — kg

Kangaroo is a next-generation bipedal humanoid unveiled in 2021, standing 160 cm and weighing 35-40 kg. It uses custom ball-screw linear actuators in a serial-parallel hybrid architecture, representing a shift from TALOS's design.

85.0/100 validated with warnings · 1/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): aiwiki.ai, github.com
✓ Citations resolve 1 of 2 cited URLs resolved
✓ Citations on-topic 1 of 1 fetched page(s) mention Kangaroo
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: 1 claim(s) supported, 5 not addressed by the evidence

2 sources: aiwiki.ai github.com View robot →

Spec update Update TALOS high confidence auto-approved

2026-08-28T04:58:37+00:00

Field	Before		After
status	research	→	commercial
onboard_compute	Intel i7, Ubuntu LTS + RTOS, EtherCAT up to 5 kHz	→	two i7 computers
battery_kwh	—	→	1.08
price_usd	—	→	900000

87.5/100 validated with warnings · 5/5 citations resolved

✓ Schema record matches the schema
✓ Source count 5 source(s) cited
✓ Source independence 5 independent domain(s): biorobot.ai, grabarobot.com, humanoidindex.org, pal-robotics.com, robolist.ai
✓ Citations resolve 5 of 5 cited URLs resolved
✓ Citations on-topic 5 of 5 fetched page(s) mention TALOS
✓ Value plausibility all numeric values within plausible ranges
! Internal consistency status 'commercial' with autonomy 'research' is contradictory
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
? Second-model check second model returned unparseable output

5 sources: humanoidindex.org www.grabarobot.com www.robolist.ai biorobot.ai View robot →

Spec update Update T800 medium confidence auto-approved

2026-08-28T04:57:07+00:00

Field	Before		After
autonomy_level	supervised-autonomy	→	research

70.0/100 validated with warnings · 1/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): eet-china.com, huxiu.com
! Citations resolve 1 of 2 cited URLs did not resolve
✓ Citations on-topic 1 of 1 fetched page(s) mention T800
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: www.eet-china.com www.huxiu.com View robot →

Spec update Update Optimus Gen 2 high confidence auto-approved

2026-08-28T04:55:14+00:00

Field	Before		After
ai_model	Tesla FSD-derived vision-only neural net	→	FSD-derived vision-only neural network
onboard_compute	FSD vision	→	Tesla FSD computer (AI4)
total_dof	28	→	40

72.5/100 validated with warnings · 3/3 citations resolved

✓ Schema record matches the schema
✓ Source count 3 source(s) cited
✓ Source independence 3 independent domain(s): biped.news, botinfo.ai, optimusk.blog
✓ Citations resolve 3 of 3 cited URLs resolved
✓ Citations on-topic 3 of 3 fetched page(s) mention Optimus Gen 2
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
! Change magnitude total_dof changes ×1.4 (28 → 40)
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

3 sources: optimusk.blog www.botinfo.ai biped.news View robot →

Company update Update XPeng auto-approved

2026-08-27T04:50:51+00:00

Field	Before		After
funding_usd_millions	—	→	900
valuation_usd_billions	—	→	6.3

100.0/100 validated · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): humanoidsdaily.com, theaiinsider.tech
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention XPeng
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
? Second-model check second model returned unparseable output

2 sources: theaiinsider.tech www.humanoidsdaily.com

Spec update Update Iron (next-gen) high confidence auto-approved

2026-08-27T04:50:51+00:00

Field	Before		After
status	prototype	→	pilot
ai_model	XPeng VLA 2.0 (Vision-Language-Action)	→	VLA2.0
onboard_compute	3x XPeng Turing AI chips (~2,250 TOPS)	→	Three Turing AI chips, 2,250 TOPS
height_cm	173	→	178
payload_kg	—	→	20
runtime_hours	—	→	4
deployments	Baosteel (industrial inspection pilot)	→	XPeng P7+ factory line (Guangzhou) - sorting, transport, screw installation (hundreds of units), Baosteel - equipment inspection, monitoring, fault detection (partnership announced), XPeng showrooms (China) - tour guide, sales assistant, reception (planned Q1 2027)

100.0/100 validated · 4/4 citations resolved

✓ Schema record matches the schema
✓ Source count 4 source(s) cited
✓ Source independence 4 independent domain(s): embodiedglobal.com, humanoidsdaily.com, interestingengineering.com, robocouture.org
✓ Citations resolve 4 of 4 cited URLs resolved
✓ Citations on-topic 4 of 4 fetched page(s) mention Iron (next-gen)
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
? Second-model check second model returned unparseable output

4 sources: www.humanoidsdaily.com interestingengineering.com robocouture.org embodiedglobal.com View robot →

Spec update Update Walker S2 high confidence auto-approved

2026-08-27T04:48:35+00:00

Field	Before		After
ai_model	BrainNet 2.0 + Co-Agents multimodal reasoning	→	BrainNet 2.0 + Co-Agent
weight_kg	73	→	95
hand_dof_per_hand	11	→	22
price_usd	90000	→	160000
deployments	BYD, Geely Auto, Zeekr, NIO, FAW-Volkswagen, Audi FAW, BAIC New Energy, Foxconn, SF Express	→	BYD, Foxconn, Airbus, Texas Instruments, Hitachi Elevator, JD Logistics, BMW, SF Express, Fangchenggang border (Vietnam border patrol, $37M contract)

72.5/100 validated with warnings · 5/6 citations resolved

✓ Schema record matches the schema
✓ Source count 14 source(s) cited
✓ Source independence 14 independent domain(s): actiontoaction.ai, aibrew.news, aibusiness.com, aiwiki.ai, humanoid.guide, ifactoryapp.com, industrysourcing.cn, robots-maker.com, robotsbeat.com, robottesters.com, sonnyrobotics.com, startuprecipe.co.kr, ubtrobot.com, wedoany.com
! Citations resolve 1 of 6 cited URLs did not resolve
✓ Citations on-topic 5 of 5 fetched page(s) mention Walker S2
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
! Change magnitude hand_dof_per_hand changes ×2.0 (11 → 22)
✓ Autonomy claim no elevated autonomy claim to challenge
? Second-model check second model returned unparseable output

14 sources: robottesters.com www.actiontoaction.ai sonnyrobotics.com robots-maker.com View robot →

Company update Update Locus Robotics auto-approved

2026-08-27T04:46:13+00:00

Field	Before		After
valuation_usd_billions	—	→	2.0
hq	Wilmington, Massachusetts, USA	→	Wilmington, MA

100.0/100 validated · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): linkedin.com, sacra.com
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Locus Robotics
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
✓ Second-model check checked independently by glm: all 2 claim(s) supported

2 sources: sacra.com www.linkedin.com

Spec update Update Optimus Gen 2 medium confidence auto-approved

2026-08-27T04:45:43+00:00

Field	Before		After
onboard_compute	Tesla FSD chip	→	FSD vision
speed_ms	—	→	2.24
deployments	Tesla Fremont factory (over 1,000 units deployed internally), Tesla Austin factory (internal deployment)	→	Tesla Fremont factory (over 1,200 units deployed internally), Tesla Austin factory (internal deployment)

100.0/100 validated · 3/3 citations resolved

✓ Schema record matches the schema
✓ Source count 3 source(s) cited
✓ Source independence 3 independent domain(s): biped.news, robohuman.org, robotwale.com
✓ Citations resolve 3 of 3 cited URLs resolved
✓ Citations on-topic 3 of 3 fetched page(s) mention Optimus Gen 2
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
? Second-model check second model returned unparseable output

3 sources: biped.news robohuman.org www.robotwale.com View robot →

Spec update Update 9RX Autonomous high confidence auto-approved

2026-08-26T04:36:35+00:00

Field	Before		After
status	limited-production	→	commercial
ai_model	—	→	Neural network for obstacle detection
onboard_compute	—	→	NVIDIA-based VPU (2 units)
weight_kg	—	→	33566
form_factor	wheeled-heavy	→	other
price_usd	—	→	1100000
deployments	Large-scale tillage on US row-crop farms	→	Gooseneck Implement (North Dakota, USA) - field testing, Cleveland Agriculture (Mungindi, QLD, Australia) - 55,000 ha wheat/chickpea planting

85.0/100 validated with warnings · 2/3 citations resolved

✓ Schema record matches the schema
✓ Source count 3 source(s) cited
✓ Source independence 3 independent domain(s): heavydutyyard.com, mondayrobotics.com, wheelerbrothers.com
! Citations resolve 1 of 3 cited URLs did not resolve
✓ Citations on-topic 2 of 2 fetched page(s) mention 9RX Autonomous
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
? Second-model check second model returned unparseable output

3 sources: www.heavydutyyard.com mondayrobotics.com wheelerbrothers.com View robot →

New model New model: Optimus Gen 2 (Tesla) high confidence auto-approved

2026-08-26T04:34:55+00:00

status pilot autonomy supervised-autonomy DoF 28 payload 20 kg deployments 2

Gen 2 remains deployed internally at Tesla factories with over 1,000 units reported, performing battery sorting and parts handling under human oversight. It has 28 body DoF and 11 DoF per hand, and is not for sale.

85.0/100 validated with warnings · 3/3 citations resolved

✓ Schema record matches the schema
✓ Source count 3 source(s) cited
✓ Source independence 2 independent domain(s): mekra.blog, optimusk.blog
✓ Citations resolve 3 of 3 cited URLs resolved
✓ Citations on-topic 3 of 3 fetched page(s) mention Optimus Gen 2
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim 'supervised-autonomy' not contradicted by the cited evidence
! Second-model check checked independently by glm: 4 claim(s) supported, 2 not addressed by the evidence

3 sources: optimusk.blog optimusk.blog mekra.blog View robot →

Spec update Update ANYmal D medium confidence auto-approved

2026-08-25T04:27:40+00:00

Field	Before		After
deployments	Northern Lights CCS facility, Norway (Equinor/TotalEnergies/Shell), PETRONAS, Shell, BASF, Vale, Siemens Energy	→	Northern Lights CCS facility, Norway (Equinor/TotalEnergies/Shell), PETRONAS, Shell, BASF, Vale, Siemens Energy, bp, Equinor, ENI, Petrobras, NVIDIA

77.5/100 validated with warnings · 1/1 citations resolved

✓ Schema record matches the schema
✓ Source count 1 source(s) cited
! Source independence all citations are on one domain (aiworld.eu) — not independent corroboration
✓ Citations resolve 1 of 1 cited URLs resolved
✓ Citations on-topic 1 of 1 fetched page(s) mention ANYmal D
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

1 sources: aiworld.eu View robot →

Company update Update ANYbotics auto-approved

2026-08-25T04:27:39+00:00

Field	Before		After
funding_usd_millions	132.0	→	152.0

92.5/100 validated with warnings · 1/1 citations resolved

✓ Schema record matches the schema
✓ Source count 1 source(s) cited
! Source independence all citations are on one domain (aiworld.eu) — not independent corroboration
✓ Citations resolve 1 of 1 cited URLs resolved
✓ Citations on-topic 1 of 1 fetched page(s) mention ANYbotics
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
✓ Second-model check checked independently by glm: all 1 claim(s) supported

1 sources: aiworld.eu

Spec update Update ANYmal X medium confidence auto-approved

2026-08-25T04:27:39+00:00

Field	Before		After
deployments	Oil, gas and chemical plants in Ex-rated zones	→	Northern Lights CCS facility, Norway
first_unveiled_year	2023	→	2026

77.5/100 validated with warnings · 1/1 citations resolved

✓ Schema record matches the schema
✓ Source count 1 source(s) cited
! Source independence all citations are on one domain (aiworld.eu) — not independent corroboration
✓ Citations resolve 1 of 1 cited URLs resolved
✓ Citations on-topic 1 of 1 fetched page(s) mention ANYmal X
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

1 sources: aiworld.eu View robot →

Company update Update 1X Technologies auto-approved

2026-08-25T04:25:00+00:00

Field	Before		After
hq	Moss, Norway	→	Palo Alto, California, USA

100.0/100 validated · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): github.com, leaderobot.com
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention 1X Technologies
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
✓ Second-model check checked independently by deepseek: all 1 claim(s) supported

2 sources: www.leaderobot.com github.com

Spec update Update Abi high confidence auto-approved

2026-08-24T04:19:09+00:00

Field	Before		After
status	limited-production	→	commercial
ai_model	Andromeda OS — generative AI/LLM stack for emotion recognition and conversation	→	ChatGPT-4, generative AI
form_factor	wheeled-base	→	wheeled-biped
deployments	22 units across mecwacare aged-care homes in Australia, supporting 1,500+ residents	→	22 units across mecwacare aged-care homes in Australia, supporting 1,500+ residents, First US deployment at Eskaton Village Carmichael in Northern California, US waitlist opened for assisted living providers

85.0/100 validated with warnings · 3/3 citations resolved

✓ Schema record matches the schema
✓ Source count 3 source(s) cited
✓ Source independence 3 independent domain(s): embodiedglobal.com, helloleaders.com.au, theaiinsider.tech
✓ Citations resolve 3 of 3 cited URLs resolved
✓ Citations on-topic 3 of 3 fetched page(s) mention Abi
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: 2 claim(s) supported, 2 not addressed by the evidence

3 sources: helloleaders.com.au theaiinsider.tech embodiedglobal.com View robot →

Company update Update Andromeda Robotics auto-approved

2026-08-24T04:19:08+00:00

Field	Before		After
valuation_usd_billions	—	→	0.1
hq	Melbourne, Australia	→	Melbourne, Victoria, Australia

100.0/100 validated · 3/3 citations resolved

✓ Schema record matches the schema
✓ Source count 3 source(s) cited
✓ Source independence 3 independent domain(s): cfotech.com.au, insideageing.com.au, onimpact.com.au
✓ Citations resolve 3 of 3 cited URLs resolved
✓ Citations on-topic 2 of 3 fetched page(s) mention Andromeda Robotics
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
? Second-model check second model returned unparseable output

3 sources: cfotech.com.au insideageing.com.au onimpact.com.au

Spec update Update SPACEO M1 medium confidence auto-approved

2026-08-24T04:16:24+00:00

Field	Before		After
status	prototype	→	commercial
ai_model	FusionMax — in-house >2B-parameter multimodal AI model	→	FusionMax
onboard_compute	Onboard Linux-based compute	→	AMD Ryzen AI Embedded X100 Series
speed_ms	1.94	→	2.0
total_dof	—	→	16
dexterous_hands	—	→	False
form_factor	wheeled-base	→	wheeled-biped
deployments	Claimed pilot relationships with Tata Motors, Indian Railways, Indian Army, Emirates Airlines (unverified independently)	→	Pilot at Pune Airport (8-10 day trial, arrival terminal), Claimed pilot relationships with Tata Motors, Indian Railways, Indian Army, Emirates Airlines (unverified independently)

100.0/100 validated · 4/4 citations resolved

✓ Schema record matches the schema
✓ Source count 4 source(s) cited
✓ Source independence 4 independent domain(s): brightwire.news, humanoid.guide, muksrobotics.com, zarintv.com
✓ Citations resolve 4 of 4 cited URLs resolved
✓ Citations on-topic 4 of 4 fetched page(s) mention SPACEO M1
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
? Second-model check second model returned unparseable output

4 sources: humanoid.guide?p=4080 muksrobotics.com zarintv.com www.brightwire.news View robot →

Spec update Update Optimus (Gen 2/3) low confidence human-approved

2026-08-23T04:44:45+00:00

Field	Before		After
status	prototype	→	pilot
deployments	Tesla Fremont factory R&D/pilot (basic tasks only, not material production as of Q1 2026)	→	Tesla Fremont factory R&D/pilot (basic tasks only, not material production as of Q1 2026), Small-batch trial production per supply-chain reports

85.0/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): github.com, smzdm.com
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 1 of 2 fetched page(s) mention Optimus (Gen 2/3)
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: post.smzdm.com github.com View robot →

Spec update Update L7 medium confidence human-approved

2026-08-23T04:44:36+00:00

Field	Before		After
status	limited-production	→	commercial
autonomy_level	supervised-autonomy	→	task-autonomous
ai_model	EraAI	→	ERA-42
deployments	10+ logistics centers with China Post and SF Express (thousand-unit scale, 2026)	→	SF Express and China Post: 10+ logistics centers nationwide (routine operation), Geely and Alibaba as deep partners, Overseas commercial services

100.0/100 validated · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): qq.com, sohu.com
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 1 of 2 fetched page(s) mention L7
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
? Second-model check second model returned unparseable output

2 sources: news.qq.com www.sohu.com View robot →

Spec update Update Moz1 medium confidence human-approved

2026-08-23T04:44:19+00:00

Field	Before		After
ai_model	Spirit v1 (VLA)	→	Spirit v1.6
form_factor	wheeled-base	→	bipedal
deployments	CATL battery production line (via 'Xiaomo' variant)	→	CATL battery production line (via 'Xiaomo' variant), Bosch China factories and logistics centers (partnership)

85.0/100 validated with warnings · 2/3 citations resolved

✓ Schema record matches the schema
✓ Source count 3 source(s) cited
✓ Source independence 3 independent domain(s): aiwiki.ai, dev.to, spirit-ai.com
✓ Citations resolve 2 of 3 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Moz1
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

3 sources: aiwiki.ai www.spirit-ai.com dev.to View robot →

Spec update Update Q5 medium confidence human-approved

2026-08-23T04:44:01+00:00

Field	Before		After
dexterous_hands	True	→	False
deployments	CES 2026 showcase	→	Retail and government/commercial service scenarios

85.0/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): qq.com, sohu.com
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 1 of 2 fetched page(s) mention Q5
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: news.qq.com www.sohu.com View robot →

Spec update Update Unitree H2 medium confidence human-approved

2026-08-23T04:32:13+00:00

85.0/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): github.com, sohu.com
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 1 of 2 fetched page(s) mention Unitree H2
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: www.sohu.com github.com View robot →

Spec update Update G1 high confidence human-approved

2026-08-23T04:32:07+00:00

Field	Before		After
autonomy_level	supervised-autonomy	→	task-autonomous
ai_model	—	→	UnifoLM, CoRe-MoE
onboard_compute	Unitree 8-core CPU; optional NVIDIA Jetson Orin (EDU, 40-100+ TOPS)	→	8-core high-performance CPU
height_cm	132	→	127.0
payload_kg	—	→	2.0
price_usd	16000	→	11700
deployments	MIT, Stanford, ETH Zurich, Research labs worldwide	→	2026 CCTV Spring Festival Gala, Guangmao Agriculture & Forestry Vocational College, Demonstration for German Chancellor Merz

85.0/100 validated with warnings · 1/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): canadadz.com, sohu.com
✓ Citations resolve 1 of 2 cited URLs resolved
✓ Citations on-topic 1 of 1 fetched page(s) mention G1
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by glm: 1 claim(s) supported, 5 not addressed by the evidence

2 sources: www.canadadz.com www.sohu.com View robot →

Spec update Update Optimus (Gen 2/3) medium confidence auto-approved (bulk)

2026-08-23T01:05:00+00:00

Field	Before		After
status	prototype	→	pilot
autonomy_level	teleoperated	→	task-autonomous
ai_model	FSD-derived neural nets; Grok (xAI) voice on Gen 3 prototypes	→	Tesla vision-based neural networks trained via end-to-end learning, using same architecture as FSD
onboard_compute	Tesla AI4 (AI5 SoC planned for Gen 3, tape-out April 2026)	→	Tesla FSD Computer (Hardware 4.0 / HW4)
payload_kg	20	→	22
total_dof	28	→	40
hand_dof_per_hand	22	→	11
deployments	Tesla Fremont factory R&D/pilot (basic tasks only, not material production as of Q1 2026)	→	Gigafactory Texas: supervised logistics and battery cell sorting, Gigafactory Nevada: supervised logistics and battery cell sorting

87.5/100 validated with warnings · 0/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 1 independent domain(s): tesla.com
? Citations resolve all 2 cited URLs blocked automated fetching (bot wall) — unverified
? Citations on-topic no page text retrieved
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
! Change magnitude total_dof changes ×1.4 (28 → 40); hand_dof_per_hand changes ÷2.0 (22 → 11)
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
? Second-model check no page text could be fetched to check against

2 sources: www.tesla.com www.tesla.com View robot →

Spec update Update Figure 01 medium confidence auto-approved (bulk)

2026-08-23T01:05:00+00:00

Field	Before		After
ai_model	OpenAI vision-language model (speech-to-speech integration)	→	OpenAI large multimodal model
funding_usd_millions	—	→	2675
valuation_usd_billions	—	→	39

92.5/100 validated with warnings · 0/3 citations resolved

✓ Schema record matches the schema
✓ Source count 3 source(s) cited
! Source independence all citations are on one domain (zhihu.com) — not independent corroboration
? Citations resolve all 3 cited URLs blocked automated fetching (bot wall) — unverified
? Citations on-topic no page text retrieved
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
? Second-model check no page text could be fetched to check against

3 sources: zhuanlan.zhihu.com zhuanlan.zhihu.com zhuanlan.zhihu.com View robot →

Spec update Update Figure 01 high confidence auto-approved (bulk)

2026-08-23T01:04:59+00:00

Field	Before		After
autonomy_level	supervised-autonomy	→	task-autonomous
ai_model	OpenAI vision-language model (speech-to-speech integration)	→	OpenAI-trained large multimodal model (vision + language) for high-level reasoning; Figure neural-network visuomotor policies mapping pixels to actions (10 Hz image input, 200 Hz action output)

92.5/100 validated with warnings · 0/4 citations resolved

✓ Schema record matches the schema
✓ Source count 4 source(s) cited
! Source independence all citations are on one domain (zhihu.com) — not independent corroboration
? Citations resolve all 4 cited URLs blocked automated fetching (bot wall) — unverified
? Citations on-topic no page text retrieved
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
? Second-model check no page text could be fetched to check against

4 sources: zhuanlan.zhihu.com zhuanlan.zhihu.com zhuanlan.zhihu.com zhuanlan.zhihu.com View robot →

Spec update Update Figure 01 medium confidence auto-approved (bulk)

2026-08-23T01:04:59+00:00

Field	Before		After
autonomy_level	supervised-autonomy	→	task-autonomous
ai_model	OpenAI vision-language model (speech-to-speech integration)	→	OpenAI-trained large multimodal model for perception, dialogue, planning, and behaviour selection; Figure neural-network visuomotor policies (pixel-to-action, 10 Hz image input → 200 Hz action output) for dexterous motion
funding_usd_millions	—	→	2675
valuation_usd_billions	—	→	39

92.5/100 validated with warnings · 0/3 citations resolved

✓ Schema record matches the schema
✓ Source count 3 source(s) cited
! Source independence all citations are on one domain (zhihu.com) — not independent corroboration
? Citations resolve all 3 cited URLs blocked automated fetching (bot wall) — unverified
? Citations on-topic no page text retrieved
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
? Second-model check no page text could be fetched to check against

3 sources: zhuanlan.zhihu.com zhuanlan.zhihu.com zhuanlan.zhihu.com View robot →

Spec update Update Unitree H2 auto-approved (bulk)

2026-08-23T01:04:59+00:00

Field	Before		After
deployments		→	2026 CCTV Spring Festival Gala (武BOT program)

100.0/100 validated · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): sina.com.cn, sohu.com
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Unitree H2
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
✓ Second-model check checked independently by deepseek: all 1 claim(s) supported

2 sources: finance.sina.com.cn www.sohu.com View robot →

Spec update Update Figure 02 high confidence auto-approved (bulk)

2026-08-23T01:04:58+00:00

Field	Before		After
status	pilot	→	retired
ai_model	Helix (in-house VLA, dual-system 7-9 Hz + 200 Hz)	→	Helix VLA (in-house, from Feb 2025)
onboard_compute	Dual NVIDIA RTX-based GPU modules (~3x Figure 01 inference)	→	Dual NVIDIA RTX GPU-based modules
payload_kg	20	→	25
deployments	BMW Group Plant Spartanburg (10-month pilot, ~1,250 hrs across 30,000+ X3 vehicles)	→	BMW Group Plant Spartanburg (11-month pilot, ended Nov 2025; 30,000+ vehicles, 90,000+ parts)

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Figure 02
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update Pepper medium confidence auto-approved (bulk)

2026-08-23T01:04:58+00:00

Field	Before		After
status	retired	→	commercial

100.0/100 validated · 2/3 citations resolved

✓ Schema record matches the schema
✓ Source count 3 source(s) cited
✓ Source independence 3 independent domain(s): gongboshi.com, robotlab.com, softbankrobotics.com
✓ Citations resolve 2 of 3 cited URLs resolved
✓ Citations on-topic 1 of 2 fetched page(s) mention Pepper
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
? Second-model check no second provider available for corroboration

3 sources: emea.softbankrobotics.com www.robotlab.com softbank-china.gongboshi.com View robot →

Spec update Update NAO medium confidence auto-approved (bulk)

2026-08-23T01:04:58+00:00

Field	Before		After
status	retired	→	commercial

87.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): robotlab.com, softbankrobotics.com
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 1 of 2 fetched page(s) mention NAO
✓ Value plausibility all numeric values within plausible ranges
! Internal consistency 0.0625 kWh over 1.5 h implies an implausible 0.04 kW draw
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
? Second-model check no second provider available for corroboration

2 sources: emea.softbankrobotics.com www.robotlab.com View robot →

Spec update Update Figure 01 high confidence auto-approved (bulk)

2026-08-23T01:04:57+00:00

Field	Before		After
autonomy_level	supervised-autonomy	→	task-autonomous
ai_model	OpenAI vision-language model (speech-to-speech integration)	→	OpenAI Large Multimodal Model + Figure Visuomotor Neural Network Policy
onboard_compute	—	→	GPU

92.5/100 validated with warnings · 0/4 citations resolved

✓ Schema record matches the schema
✓ Source count 4 source(s) cited
! Source independence all citations are on one domain (zhihu.com) — not independent corroboration
? Citations resolve all 4 cited URLs blocked automated fetching (bot wall) — unverified
? Citations on-topic no page text retrieved
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
? Second-model check no page text could be fetched to check against

4 sources: zhuanlan.zhihu.com zhuanlan.zhihu.com zhuanlan.zhihu.com zhuanlan.zhihu.com View robot →

Spec update Update Calvin-40 high confidence auto-approved (bulk)

2026-08-23T01:04:57+00:00

Field	Before		After
status	pilot	→	commercial
autonomy_level	supervised-autonomy	→	task-autonomous
ai_model	RL 3D vision/recognition stack; balance control carried over from Wandercraft's exoskeleton platform	→	NVIDIA Isaac GR00T (N1 foundation model) & Physical AI / VLM
onboard_compute	—	→	NVIDIA Jetson edge AI platform / NVIDIA RTX GPU compute
total_dof	32	→	28
deployments	Renault Group Douai plant pilot (350 units planned across ~10 European sites by 2026-2027)	→	Renault Group Douai plant pilot (350 units planned across ~10 European sites by 2026-2027), SAPA (Italian automotive Tier 1 supplier) manufacturing facilities

92.5/100 validated with warnings · 5/6 citations resolved

✓ Schema record matches the schema
✓ Source count 14 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 5 of 6 cited URLs resolved
✓ Citations on-topic 4 of 5 fetched page(s) mention Calvin-40
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
? Second-model check second model returned unparseable output

14 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update THEMIS V2 high confidence auto-approved (bulk)

2026-08-23T01:04:57+00:00

Field	Before		After
status	research	→	commercial
autonomy_level	research	→	task-autonomous
ai_model	ROS-based control with LLM-assisted task planning	→	ROS-based Humanoid OS
onboard_compute	~200 TOPS AI compute (unspecified chip)	→	200 TOPS AI computing power
weight_kg	36	→	36.5
payload_kg	—	→	15.0
deployments		→	Academic research laboratories, Research universities

77.5/100 validated with warnings · 1/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 1 of 2 cited URLs resolved
✓ Citations on-topic 1 of 1 fetched page(s) mention THEMIS V2
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by deepseek: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update Figure 03 high confidence auto-approved (bulk)

2026-08-23T01:04:57+00:00

Field	Before		After
status	prototype	→	commercial
autonomy_level	supervised-autonomy	→	task-autonomous
ai_model	Helix (vision-language-action model)	→	Helix 02 VLA (System 0)
onboard_compute	dual low-power embedded GPUs	→	Onboard low-power GPUs; 10 Gbps mmWave data offload
height_cm	168	→	173
weight_kg	60	→	64
total_dof	—	→	35
hand_dof_per_hand	—	→	16
deployments		→	BMW Group Plant Spartanburg (logistics sequencing, deployed June 2026), Catalyst Brands Reno, NV Distribution Logistics Center (sorting & packing, signed May 2026), Brookfield Asset Management (Project Go-Big, residential data collection)

77.5/100 validated with warnings · 2/3 citations resolved

✓ Schema record matches the schema
✓ Source count 3 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 3 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Figure 03
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

3 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update H2 Plus high confidence auto-approved (bulk)

2026-08-23T01:04:56+00:00

Field	Before		After
status	research	→	commercial
autonomy_level	research	→	task-autonomous
onboard_compute	NVIDIA Jetson Thor (Blackwell GPU, 128GB unified memory, 2070 TFLOPS)	→	NVIDIA Jetson Thor
price_usd	—	→	100000
deployments	Academic units to Ai2, ETH Zurich, Stanford Robotics Center, UC San Diego (late 2026)	→	NVIDIA Isaac GR00T Reference Platform, Ai2, ETH Zurich, Stanford Robotics Center, UC San Diego

77.5/100 validated with warnings · 1/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 1 of 2 cited URLs resolved
✓ Citations on-topic 1 of 1 fetched page(s) mention H2 Plus
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by deepseek: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update ARMAR-7 high confidence auto-approved (bulk)

2026-08-23T01:04:56+00:00

Field	Before		After
ai_model	ArmarX cognitive software architecture; learning-from-human-demonstration	→	ArmarX; LLM agents (ROSA, AutoGPT+P), Episodic Memory Verbalization, imitation learning
onboard_compute	—	→	Multi-PC cluster running Linux; hands with FPGAs for on-finger CNN inference
payload_kg	—	→	2
deployments		→	Hannover Messe 2026 (April 2026), automatica 2025 (Munich, June 2025), 1st euROBIN Coopetition (Nancy, November 2024), Robotics Institute Germany (RIG) demonstrator

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention ARMAR-7
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update Agile ONE high confidence auto-approved (bulk)

2026-08-23T01:04:56+00:00

Field	Before		After
status	prototype	→	pilot
autonomy_level	supervised-autonomy	→	task-autonomous
ai_model	Agile Robots' layered 'Physical AI' architecture	→	AgileCore / Google DeepMind Gemini Robotics 2
onboard_compute	—	→	Multi-tiered edge compute architecture
deployments		→	Agile Robots Kaufbeuren Manufacturing Hub, Hannover Messe 2026, NVIDIA GTC 2026

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Agile ONE
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by deepseek: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update EVE high confidence auto-approved (bulk)

2026-08-23T01:04:56+00:00

Field	Before		After
ai_model	—	→	Embodied AI with GPT-4
onboard_compute	Intel i7 + NVIDIA Jetson Xavier	→	Intel Core i7 + NVIDIA Jetson Xavier
height_cm	183	→	188
weight_kg	83	→	87
hand_dof_per_hand	0	→	1
first_unveiled_year	2018	→	2022

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 1 of 2 fetched page(s) mention EVE
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update KAIST Humanoid v0.7 medium confidence auto-approved (bulk)

2026-08-23T01:04:55+00:00

Field	Before		After
autonomy_level	research	→	task-autonomous
ai_model	RL locomotion policies trained in simulation, transferred to hardware	→	Deep Reinforcement Learning (DRL) with human motion demonstration priors and Sim2Real transfer
speed_ms	3.6	→	3.61
dexterous_hands	—	→	False

77.5/100 validated with warnings · 1/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 1 of 2 cited URLs resolved
✓ Citations on-topic 1 of 1 fetched page(s) mention KAIST Humanoid v0.7
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update Jupiter high confidence auto-approved (bulk)

2026-08-23T01:04:55+00:00

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 1 of 2 fetched page(s) mention Jupiter
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by deepseek: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update Kuavo-5 high confidence auto-approved (bulk)

2026-08-23T01:04:55+00:00

Field	Before		After
autonomy_level	teleoperated	→	task-autonomous
ai_model	—	→	Huawei Pangu Embodied AI Model
onboard_compute	—	→	RK3588 + NVIDIA Orin NX (157 TOPS)
payload_kg	14	→	20
deployments	Universities, exhibition halls, automotive plants (China), Torchbearer at China's 15th National Games, Nov 2025	→	NIO EV Smart Factory, China Southern Power Grid & China Mobile (1,200 km 5G-A substation inspection), FAW Group, Haier Group, Jiangsu Hengtong

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Kuavo-5
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update MIRO U medium confidence auto-approved (bulk)

2026-08-23T01:04:55+00:00

Field	Before		After
status	prototype	→	pilot
autonomy_level	supervised-autonomy	→	task-autonomous
ai_model	—	→	Midea Factory Brain multi-agent architecture and industrial multi-modal LLM/VLM
onboard_compute	—	→	Midea proprietary AI edge industrial controller
form_factor	other	→	wheeled-base
deployments	Midea Wuxi high-end washing machine factory (planned end of 2025)	→	Midea High-End Washing Machine Factory (Wuxi, Jiangsu Province) - deployed December 2025

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 1 of 2 fetched page(s) mention MIRO U
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update Unitree H2 high confidence auto-approved (bulk)

2026-08-23T01:04:54+00:00

Field	Before		After
deployments		→	2026 CCTV Spring Festival Gala

85.0/100 validated with warnings · 3/3 citations resolved

✓ Schema record matches the schema
✓ Source count 3 source(s) cited
✓ Source independence 2 independent domain(s): sina.com.cn, sohu.com
✓ Citations resolve 3 of 3 cited URLs resolved
✓ Citations on-topic 3 of 3 fetched page(s) mention Unitree H2
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by deepseek: the cited evidence does not address any of the claims

3 sources: finance.sina.com.cn www.sohu.com www.sohu.com View robot →

Spec update Update G1 medium confidence auto-approved (bulk)

2026-08-23T01:04:54+00:00

Field	Before		After
autonomy_level	supervised-autonomy	→	teleoperated
ai_model	Hybrid AI + teleoperation via GITAI's 'H1' manipulation system	→	Proprietary GITAI control suite
onboard_compute	—	→	In-house proprietary real-time control hardware
height_cm	—	→	200
weight_kg	—	→	120
total_dof	—	→	21
hand_dof_per_hand	—	→	8
form_factor	arm-only	→	other
deployments	ISS technology demonstrations, JAXA joint research agreement	→	JAXA Kibo mock-up module, Simulated lunar environments, ISS technology demonstrations, JAXA joint research agreement
first_unveiled_year	2021	→	2020

70.0/100 validated with warnings · 2/3 citations resolved

✓ Schema record matches the schema
✓ Source count 3 source(s) cited
✓ Source independence 2 independent domain(s): gitai.tech, youtube.com
! Citations resolve 1 of 3 cited URLs did not resolve
✓ Citations on-topic 1 of 2 fetched page(s) mention G1
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

3 sources: gitai.tech www.youtube.com gitai.tech View robot →

Spec update Update NEO (Gamma) high confidence auto-approved (bulk)

2026-08-23T01:04:54+00:00

Field	Before		After
autonomy_level	teleoperated	→	supervised-autonomy
ai_model	1X Redwood / in-house '1X Mind'	→	Redwood AI VLA + 1X World Model
onboard_compute	1X Cortex (NVIDIA Jetson Thor)	→	NVIDIA Jetson Thor
height_cm	165	→	168
total_dof	—	→	75
hand_dof_per_hand	22	→	25
deployments	EQT (up to 10,000 units across portfolio, 2026-2030), Consumer pre-orders (US, 2026)	→	EQT (up to 10,000 units across portfolio, 2026-2030), Consumer pre-orders (10,000+ reservations, late 2026 US deliveries)

65.0/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 1 of 2 fetched page(s) mention NEO (Gamma)
✓ Value plausibility all numeric values within plausible ranges
! Internal consistency payload (25 kg) is unusually high for a 30 kg robot
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'supervised-autonomy' not contradicted by the cited evidence
! Second-model check checked independently by deepseek: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update Figure 01 high confidence auto-approved (bulk)

2026-08-23T01:04:54+00:00

Field	Before		After
autonomy_level	supervised-autonomy	→	task-autonomous
ai_model	OpenAI vision-language model (speech-to-speech integration)	→	OpenAI large multimodal model (vision + language) for cognition; Figure's own neural-network visuomotor policies for action
onboard_compute	—	→	GPU (specific chipset not stated)
first_unveiled_year	2023	→	2024

92.5/100 validated with warnings · 0/6 citations resolved

✓ Schema record matches the schema
✓ Source count 6 source(s) cited
! Source independence all citations are on one domain (zhihu.com) — not independent corroboration
? Citations resolve all 6 cited URLs blocked automated fetching (bot wall) — unverified
? Citations on-topic no page text retrieved
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
? Second-model check no page text could be fetched to check against

6 sources: zhuanlan.zhihu.com zhuanlan.zhihu.com zhuanlan.zhihu.com zhuanlan.zhihu.com View robot →

Spec update Update H1 high confidence auto-approved (bulk)

2026-08-23T01:04:53+00:00

Field	Before		After
ai_model	—	→	Model Predictive Control (MPC) + Sim2Real Reinforcement Learning
onboard_compute	Intel Core i5 + i7; optional NVIDIA Jetson Orin NX	→	Dual Intel Core i7 + NVIDIA Jetson-class upgrade path
payload_kg	—	→	10
hand_dof_per_hand	—	→	0
price_usd	99900	→	90000
deployments	MIT, Stanford, ETH Zurich, Research institutions worldwide	→	Geely (automotive manufacturing trials), Stanford University (HumanPlus research), Carnegie Mellon University (H2O project), UT Health Science Center, 16-robot troupe on 2025 CCTV Spring Festival Gala

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Unitree H1
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update Atlas (electric) high confidence auto-approved (bulk)

2026-08-23T01:04:53+00:00

Field	Before		After
status	limited-production	→	pilot
autonomy_level	supervised-autonomy	→	task-autonomous
ai_model	NVIDIA Isaac GR00T + Google DeepMind Gemini Robotics; Toyota Research Institute LBM	→	Strategic embodied AI foundation models co-developed with Google DeepMind (Gemini-based reasoning AI) + Boston Dynamics Orbit software platform
onboard_compute	NVIDIA Jetson Thor	→	Embedded high-performance robotics processors (integrated with NVIDIA simulation architecture and BD Orbit platform)
valuation_usd_billions	1.1	→	20
deployments	Hyundai HMGMA (Georgia metaplant, automotive assembly), Google DeepMind (AI research)	→	Hyundai Motor Group – Hyundai Metaplant America (HMGMA, Georgia, USA) for automotive parts sequencing and material handling, Hyundai Motor Group – Robotics Metaplant Application Center (RMAC), Google DeepMind – Research facilities for physical AI and foundation model training

65.0/100 validated with warnings · 5/5 citations resolved

✓ Schema record matches the schema
✓ Source count 5 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 5 of 5 cited URLs resolved
✓ Citations on-topic 4 of 5 fetched page(s) mention Atlas (Electric)
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
! Change magnitude status moves backwards: limited-production → pilot; valuation_usd_billions changes ×18.2 (1.1 → 20)
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by deepseek: 1 claim(s) supported, 5 not addressed by the evidence

5 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update Figure 02 high confidence auto-approved (bulk)

2026-08-23T01:04:53+00:00

Field	Before		After
status	pilot	→	retired
ai_model	Helix (in-house VLA, dual-system 7-9 Hz + 200 Hz)	→	Helix VLA / Helix 02
onboard_compute	Dual NVIDIA RTX-based GPU modules (~3x Figure 01 inference)	→	Torso-integrated Dual NVIDIA GPUs
deployments	BMW Group Plant Spartanburg (10-month pilot, ~1,250 hrs across 30,000+ X3 vehicles)	→	BMW Group Plant Spartanburg (11-month pilot completed Nov 2025, 1,250 hrs across 30,000+ X3 vehicles, 90,000+ sheet metal parts handled)

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Figure 02
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by deepseek: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update Figure 01 high confidence auto-approved (bulk)

2026-08-23T01:04:53+00:00

Field	Before		After
status	prototype	→	retired
autonomy_level	supervised-autonomy	→	task-autonomous
ai_model	OpenAI vision-language model (speech-to-speech integration)	→	OpenAI VLM/LLM (terminated Feb 2025)
onboard_compute	—	→	Custom onboard compute
hand_dof_per_hand	—	→	6
funding_usd_millions	—	→	2675
valuation_usd_billions	—	→	39
deployments		→	BMW Group Plant Spartanburg (initial 2024 proof-of-concept pilot)

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Figure 01
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by deepseek: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update Unitree R1 high confidence auto-approved (bulk)

2026-08-23T01:04:52+00:00

Field	Before		After
autonomy_level	research	→	supervised-autonomy
ai_model	onboard multimodal model (voice + image recognition)	→	Unitree UnifoLM multimodal large model
onboard_compute	8-core CPU/GPU	→	8-core CPU (base); optional NVIDIA Jetson Orin (10-100 TOPS) on EDU
height_cm	121	→	123
weight_kg	25	→	29
payload_kg	—	→	2
hand_dof_per_hand	—	→	0
deployments		→	STEM/K-12 education, university AI labs, light electronics assembly research
first_unveiled_year	2025	→	2026

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Unitree R1
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'supervised-autonomy' not contradicted by the cited evidence
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update Unitree H2 high confidence auto-approved (bulk)

2026-08-23T01:04:52+00:00

Field	Before		After
status	prototype	→	commercial
autonomy_level	research	→	supervised-autonomy
ai_model	—	→	Embodied AI models, MPC + Reinforcement Learning
onboard_compute	Intel Core i5 + i7 (optional NVIDIA Jetson AGX Thor on EDU variant)	→	Dual PC architecture (Intel Core i5 platform PC + Intel Core i7 user development PC) + configurable high-power GPU options
height_cm	180	→	182
hand_dof_per_hand	—	→	0
dexterous_hands	True	→	False
price_usd	—	→	29900
deployments		→	2026 Spring Festival Gala (sword dance performance), early industrial pilot evaluations

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Unitree H2
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'supervised-autonomy' not contradicted by the cited evidence
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update H2 Plus high confidence auto-approved (bulk)

2026-08-23T01:04:52+00:00

Field	Before		After
status	research	→	commercial
autonomy_level	research	→	supervised-autonomy
ai_model	NVIDIA Isaac GR00T	→	NVIDIA Isaac GR00T foundation models
onboard_compute	NVIDIA Jetson Thor (Blackwell GPU, 128GB unified memory, 2070 TFLOPS)	→	NVIDIA Jetson AGX Thor T5000 (Blackwell GPU, 2,070 FP4 TFLOPS, 14-core Arm CPU, 128GB unified memory) + Intel i5/i7 PCs
payload_kg	15	→	7
price_usd	—	→	100000
deployments	Academic units to Ai2, ETH Zurich, Stanford Robotics Center, UC San Diego (late 2026)	→	Academic/Enterprise reference platform for NVIDIA Isaac GR00T developers

65.0/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Unitree H2 Plus
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
! Change magnitude payload_kg changes ÷2.1 (15 → 7)
✓ Autonomy claim 'supervised-autonomy' not contradicted by the cited evidence
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update G1 high confidence auto-approved (bulk)

2026-08-23T01:04:52+00:00

Field	Before		After
ai_model	—	→	Unitree UnifoLM multimodal model
onboard_compute	Unitree 8-core CPU; optional NVIDIA Jetson Orin (EDU, 40-100+ TOPS)	→	Base: 8-core CPU; EDU: NVIDIA Jetson Orin NX (40-100 TOPS)
payload_kg	—	→	3
total_dof	23	→	43
price_usd	16000	→	13500
deployments	MIT, Stanford, ETH Zurich, Research labs worldwide	→	Japan Airlines (baggage handling trials at Haneda Airport), YY Group (facility cleaning), US Sheriff's Department, Amazon, Stanford University, MIT, 2026 Spring Festival Gala

65.0/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Unitree G1
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
! Change magnitude total_dof changes ×1.9 (23 → 43)
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update Galbot S1 medium confidence auto-approved (bulk)

2026-08-23T01:04:51+00:00

85.0/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): qq.com, sohu.com
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Galbot S1
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: m.sohu.com news.qq.com View robot →

Spec update Update Forerunner K2 medium confidence auto-approved (bulk)

2026-08-23T01:04:51+00:00

85.0/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): gotokepler.com, prnewswire.com
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Forerunner K2
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: www.gotokepler.com www.prnewswire.com View robot →

Spec update Update Forerunner K1 medium confidence auto-approved (bulk)

2026-08-23T01:04:51+00:00

85.0/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): gotokepler.com, prnewswire.com
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Forerunner K1
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: www.gotokepler.com www.prnewswire.com View robot →

Spec update Update EVE high confidence auto-approved (bulk)

2026-08-23T01:04:51+00:00

Field	Before		After
status	pilot	→	commercial
ai_model	—	→	Redwood AI architecture running ROS2 and OpenCV vision pipelines
onboard_compute	Intel i7 + NVIDIA Jetson Xavier	→	Intel Core i7 + NVIDIA Jetson Xavier
weight_kg	83	→	85
hand_dof_per_hand	0	→	1
funding_usd_millions	136.5	→	125
valuation_usd_billions	0.82	→	10
deployments	ADT Commercial / Everon night-guarding security pilots (US facilities)	→	ADT Commercial / Everon night-guarding security pilots, Global industrial facilities and security patrolling sites, Warehousing operations and healthcare facility logistics pilots
first_unveiled_year	2018	→	2020

50.0/100 validated with warnings · 1/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
! Citations resolve 1 of 2 cited URLs did not resolve
✓ Citations on-topic 1 of 1 fetched page(s) mention EVE
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
! Change magnitude valuation_usd_billions changes ×12.2 (0.82 → 10)
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by deepseek: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update Figure 01 high confidence auto-approved (bulk)

2026-08-23T01:04:50+00:00

Field	Before		After
autonomy_level	supervised-autonomy	→	task-autonomous
ai_model	OpenAI vision-language model (speech-to-speech integration)	→	OpenAI multimodal model + Figure visuomotor neural network policy
onboard_compute	—	→	Onboard GPU
funding_usd_millions	—	→	2675
valuation_usd_billions	—	→	39

92.5/100 validated with warnings · 0/4 citations resolved

✓ Schema record matches the schema
✓ Source count 4 source(s) cited
! Source independence all citations are on one domain (zhihu.com) — not independent corroboration
? Citations resolve all 4 cited URLs blocked automated fetching (bot wall) — unverified
? Citations on-topic no page text retrieved
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
? Second-model check no page text could be fetched to check against

4 sources: zhuanlan.zhihu.com zhuanlan.zhihu.com zhuanlan.zhihu.com zhuanlan.zhihu.com View robot →

Spec update Update Figure 01 medium confidence auto-approved (bulk)

2026-08-23T01:04:50+00:00

Field	Before		After
ai_model	OpenAI vision-language model (speech-to-speech integration)	→	OpenAI multimodal large model for visual reasoning and language understanding, paired with Figure neural networks for visuomotor control

92.5/100 validated with warnings · 0/3 citations resolved

✓ Schema record matches the schema
✓ Source count 3 source(s) cited
! Source independence all citations are on one domain (zhihu.com) — not independent corroboration
? Citations resolve all 3 cited URLs blocked automated fetching (bot wall) — unverified
? Citations on-topic no page text retrieved
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
? Second-model check no page text could be fetched to check against

3 sources: zhuanlan.zhihu.com zhuanlan.zhihu.com zhuanlan.zhihu.com View robot →

Spec update Update Figure 01 high confidence auto-approved (bulk)

2026-08-23T01:04:50+00:00

Field	Before		After
autonomy_level	supervised-autonomy	→	task-autonomous
ai_model	OpenAI vision-language model (speech-to-speech integration)	→	OpenAI multimodal model
height_cm	168	→	167.64

92.5/100 validated with warnings · 0/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (zhihu.com) — not independent corroboration
? Citations resolve all 2 cited URLs blocked automated fetching (bot wall) — unverified
? Citations on-topic no page text retrieved
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
? Second-model check no page text could be fetched to check against

2 sources: zhuanlan.zhihu.com zhuanlan.zhihu.com View robot →

Spec update Update Galbot G1 high confidence auto-approved (bulk)

2026-08-23T01:04:50+00:00

Field	Before		After
status	pilot	→	commercial
ai_model	Large-model general-purpose manipulation (embodied AI)	→	AstraBrain
dexterous_hands	False	→	True
form_factor	wheeled-base	→	bipedal
price_usd	90000	→	98000
deployments	Galbot Store (autonomous retail, 30+ cities), Manufacturing, Logistics, Healthcare	→	FamilyMart (Beijing), Galbot Store (autonomous retail, 30+ cities), Manufacturing, Logistics, Healthcare, 2026 CCTV Spring Festival Gala

100.0/100 validated · 3/3 citations resolved

✓ Schema record matches the schema
✓ Source count 3 source(s) cited
✓ Source independence 2 independent domain(s): qq.com, sohu.com
✓ Citations resolve 3 of 3 cited URLs resolved
✓ Citations on-topic 3 of 3 fetched page(s) mention Galbot G1
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
? Second-model check second model returned unparseable output

3 sources: m.sohu.com news.qq.com www.sohu.com View robot →

Spec update Update XR4 (Xiaozi) medium confidence auto-approved (bulk)

2026-08-23T01:04:49+00:00

Field	Before		After
status	limited-production	→	pilot
ai_model	RobotGPT (in-house multimodal LLM) on HARIX Cloud AI Robot OS	→	RobotGPT with HARIX OS
onboard_compute	—	→	Robot Control Unit (RCU) with 5G connectivity
weight_kg	75	→	67.5
payload_kg	—	→	7.5
total_dof	50	→	60
hand_dof_per_hand	—	→	6
dexterous_hands	—	→	True
actuation	electric	→	hybrid
deployments		→	Research exhibitions and industrial pilot demonstrations at WAIC testbeds
first_unveiled_year	2024	→	2023

65.0/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention XR4 (Xiaozi)
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
! Change magnitude status moves backwards: limited-production → pilot
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: 1 claim(s) supported, 5 not addressed by the evidence

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update Robody medium confidence auto-approved (bulk)

2026-08-23T01:04:49+00:00

Field	Before		After
autonomy_level	teleoperated	→	supervised-autonomy
ai_model	AI automation for routine tasks combined with human VR teleoperation over 5G	→	Proprietary physical AI and imitation learning trained on real-world teleoperation episodes
onboard_compute	—	→	NVIDIA Jetson Orin + Jetson Orin Nano
height_cm	—	→	165
weight_kg	—	→	60
total_dof	28	→	17
hand_dof_per_hand	—	→	1
dexterous_hands	—	→	False
deployments	Charite Berlin, University of Oxford, Diakonie (Germany)	→	Charité Berlin, LMU Klinikum, Diakonie, University of Oxford, Platten Meilen (Switzerland), Deutsches Museum Munich, Private home pilots (Germany/Switzerland)
first_unveiled_year	2024	→	2023

65.0/100 validated with warnings · 3/4 citations resolved

✓ Schema record matches the schema
✓ Source count 4 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 3 of 4 cited URLs resolved
✓ Citations on-topic 3 of 3 fetched page(s) mention Robody (Gen 3 / Robody Cares)
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
! Change magnitude total_dof changes ÷1.6 (28 → 17)
✓ Autonomy claim 'supervised-autonomy' not contradicted by the cited evidence
! Second-model check checked independently by deepseek: the cited evidence does not address any of the claims

4 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update NEO (Gamma) high confidence auto-approved (bulk)

2026-08-23T01:04:49+00:00

Field	Before		After
autonomy_level	teleoperated	→	supervised-autonomy
ai_model	1X Redwood / in-house '1X Mind'	→	Redwood AI (160M-param VLA) + 1X World Model
onboard_compute	1X Cortex (NVIDIA Jetson Thor)	→	NVIDIA Jetson Thor
height_cm	165	→	168
total_dof	—	→	75
hand_dof_per_hand	22	→	25
deployments	EQT (up to 10,000 units across portfolio, 2026-2030), Consumer pre-orders (US, 2026)	→	EQT portfolio (up to 10,000 units, 2026-2030), Consumer pre-orders (US, 2026), Internal Hayward factory logistics
first_unveiled_year	2024	→	2023

65.0/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention NEO (Gamma)
✓ Value plausibility all numeric values within plausible ranges
! Internal consistency payload (25 kg) is unusually high for a 30 kg robot
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'supervised-autonomy' not contradicted by the cited evidence
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update EVE high confidence auto-approved (bulk)

2026-08-23T01:04:49+00:00

Field	Before		After
status	pilot	→	commercial
ai_model	—	→	Redwood VLA + imitation learning
onboard_compute	Intel i7 + NVIDIA Jetson Xavier	→	Intel Core i7 + NVIDIA Jetson Xavier
height_cm	183	→	188
weight_kg	83	→	87
hand_dof_per_hand	0	→	1
deployments	ADT Commercial / Everon night-guarding security pilots (US facilities)	→	ADT Commercial / Everon (140-unit security patrol), Sunnaas Rehabilitation Hospital (Norway), Internal 1X data-gathering fleets

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention EVE
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update NEO (Gamma) medium confidence auto-approved (bulk)

2026-08-23T01:04:48+00:00

Field	Before		After
deployments	EQT (up to 10,000 units across portfolio, 2026-2030), Consumer pre-orders (US, 2026)	→	EQT (up to 10,000 units across portfolio, 2026-2030), Consumer pre-orders (US, 2026), Early access programs in hundreds of homes

72.5/100 validated with warnings · 3/3 citations resolved

✓ Schema record matches the schema
✓ Source count 3 source(s) cited
✓ Source independence 3 independent domain(s): capitalispartners.com, github.com, zerogtalent.com
✓ Citations resolve 3 of 3 cited URLs resolved
✓ Citations on-topic 1 of 3 fetched page(s) mention NEO (Gamma)
✓ Value plausibility all numeric values within plausible ranges
! Internal consistency payload (25 kg) is unusually high for a 30 kg robot
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

3 sources: zerogtalent.com capitalispartners.com github.com View robot →

Spec update Update AlphaBot 2 high confidence auto-approved (bulk)

2026-08-23T01:04:48+00:00

Field	Before		After
status	pilot	→	commercial
autonomy_level	supervised-autonomy	→	task-autonomous
ai_model	AlphaBrain / GOVLA (Global and Omni-body Vision-Language-Action)	→	Alpha Brain (GOVLA)
onboard_compute	—	→	NVIDIA AGX Orin 64GB
height_cm	—	→	180
weight_kg	—	→	135
payload_kg	—	→	8
speed_ms	—	→	1.5
runtime_hours	—	→	6
total_dof	34	→	22
form_factor	wheeled-base	→	wheeled-biped
deployments	Automotive, semiconductor, biomanufacturing, public service, retail pilots	→	Dongfeng Liuzhou Motor Co. automotive factory (since June 2025), HKC Corporation semiconductor display manufacturing (over 1,000 units over 3 years), Airports in Chinese first-tier cities (Q3 2025), Public service and retail locations across 10+ provinces in China

87.5/100 validated with warnings · 5/5 citations resolved

✓ Schema record matches the schema
✓ Source count 5 source(s) cited
✓ Source independence 5 independent domain(s): aparobot.com, futurum.media, interestingengineering.com, robothub.app, toutiao.com
✓ Citations resolve 5 of 5 cited URLs resolved
✓ Citations on-topic 4 of 5 fetched page(s) mention AlphaBot 2
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
! Change magnitude total_dof changes ÷1.5 (34 → 22)
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
? Second-model check second model returned unparseable output

5 sources: www.toutiao.com www.robothub.app www.aparobot.com futurum.media View robot →

New model New model: Mipa (NEURA Robotics) low confidence auto-approved (bulk)

2026-08-23T01:04:48+00:00

status prototype autonomy research DoF — payload — kg

Mipa is a home robot mentioned in the context of NEURA's showcase at IFA, indicating a new product line for domestic assistance. Details are limited, but it represents an expansion beyond industrial humanoids.

85.0/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): neura-robotics.com, qq.com
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Mipa
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: 1 claim(s) supported, 2 not addressed by the evidence

2 sources: new.qq.com neura-robotics.com View robot →

New model New model: ARMAR-6 (KIT High Performance Humanoid Technologies Lab) high confidence auto-approved (bulk)

2026-08-23T01:04:47+00:00

status research autonomy task-autonomous DoF 28 payload 10 kg deployments 2

ARMAR-6 is an industrial collaborative maintenance assistant with 28 DoF and 10 kg payload per arm, tested in the EU SecondHands project. It remains an active research platform at KIT.

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention ARMAR-6
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

New model New model: ARMAR-DE (KIT High Performance Humanoid Technologies Lab) high confidence auto-approved (bulk)

2026-08-23T01:04:47+00:00

status research autonomy task-autonomous DoF 32 payload 10 kg deployments 2

ARMAR-DE is designed for hazardous environments and decontamination tasks, used in the ROBDEKON project. It features torque-controlled SAC units and LLM-based planning.

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention ARMAR-DE
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

New model New model: Mila (Midea Group) medium confidence auto-approved (bulk)

2026-08-23T01:04:46+00:00

status prototype autonomy task-autonomous DoF 40 payload — kg deployments 1

Mila is Midea's commercial and household service humanoid, demonstrated at WAIC 2025 and IFA 2025, capable of operating home appliances and scheduled for deployment in Midea experience stores in 2026.

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 1 of 2 fetched page(s) mention Mila
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

New model New model: Cloud Ginger 2.0 (XR-2) (Dataa Robotics (CloudMinds)) high confidence auto-approved (bulk)

2026-08-23T01:04:46+00:00

status commercial autonomy task-autonomous DoF 41 payload 5 kg deployments 5

Cloud Ginger 2.0 is a wheeled service humanoid with 41 DoF and 5 kg arm payload, designed for customer interaction and light manipulation. It is commercially deployed in healthcare, hospitality, and retail settings.

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Cloud Ginger 2.0 (XR-2)
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

New model New model: AELOS Series (Leju Robot) high confidence auto-approved (bulk)

2026-08-23T01:04:45+00:00

status commercial autonomy teleoperated DoF 22 payload 0.5 kg deployments 3

Miniature educational humanoids used in STEM education and public exhibitions, with over 100 units shipped globally.

62.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 1 of 2 fetched page(s) mention AELOS Series
! Value plausibility height_cm=34.6 unusual (typical 40–220); weight_kg=1.5 unusual (typical 2–200)
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

New model New model: Mero X (Midea Group) medium confidence auto-approved (bulk)

2026-08-23T01:04:45+00:00

status prototype autonomy task-autonomous DoF 40 payload — kg deployments 1

Mero X is Midea's second-generation full bipedal humanoid with 40+ body joints and 20 DoF hands, designed for dynamic balance and complex terrain traversal, currently in R&D at Midea's Blue Orange Laboratory.

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 1 of 2 fetched page(s) mention Mero X
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

New model New model: Mero 1 (Midea Group) medium confidence auto-approved (bulk)

2026-08-23T01:04:45+00:00

status pilot autonomy task-autonomous DoF — payload 8 kg deployments 1

Mero 1 is Midea's first-generation industrial wheeled humanoid, deployed in a Jingzhou factory for inspection, parts transport, and assembly support, with hands rated to lift 8 kg.

92.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 1 of 2 fetched page(s) mention Mero 1
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
? Second-model check second model returned unparseable output

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

New model New model: Kuavo 1.0 (Leju Robot) high confidence auto-approved (bulk)

2026-08-23T01:04:44+00:00

status commercial autonomy supervised-autonomy DoF 26 payload 5 kg deployments 3

Original bipedal platform, first Chinese humanoid running OpenHarmony, now superseded by newer models.

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Kuavo 1.0
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim 'supervised-autonomy' not contradicted by the cited evidence
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

New model New model: ROBAN 2 (Leju Robot) medium confidence auto-approved (bulk)

2026-08-23T01:04:44+00:00

status pilot autonomy supervised-autonomy DoF 26 payload 9 kg deployments 2

Medium-sized research platform for education and workforce simulation, entering pilot production in 2026.

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 1 of 2 fetched page(s) mention ROBAN 2
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim 'supervised-autonomy' not contradicted by the cited evidence
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

New model New model: Booster K2 (Booster Robotics) medium confidence auto-approved (bulk)

2026-08-23T01:04:43+00:00

status limited-production autonomy task-autonomous DoF 22 payload — kg deployments 2

Next-gen compact humanoid, successor to K1, demonstrated mid-2026; early commercial rollout via distributors.

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Booster K2
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

New model New model: Kuavo 4Pro (Leju Robot) high confidence auto-approved (bulk)

2026-08-23T01:04:43+00:00

status commercial autonomy task-autonomous DoF 41 payload 5 kg deployments 4

Full-size research and service humanoid with high-dynamics mobility, used for dancing and showroom guidance.

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Kuavo 4Pro
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by glm: 1 claim(s) supported, 5 not addressed by the evidence

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

New model New model: KAIST Humanoid v0.5 (KAIST Dynamic Robot Control & Design Lab) medium confidence auto-approved (bulk)

2026-08-23T01:04:42+00:00

status research autonomy task-autonomous DoF — payload — kg

KAIST Humanoid v0.5 was a precursor bipedal platform showcased in September 2025, demonstrating reinforcement learning-based locomotion. It was succeeded by v0.7 in 2026.

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention KAIST Humanoid v0.5
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

New model New model: Booster K1 (Booster Robotics) high confidence auto-approved (bulk)

2026-08-23T01:04:42+00:00

status commercial autonomy task-autonomous DoF 22 payload 1 kg deployments 4

Compact 95cm humanoid with 22 DoF, launched Oct 2025; over 1,000 units shipped across 20+ countries; RoboCup KidSize 2025 champion.

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Booster K1
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by glm: 2 claim(s) supported, 4 not addressed by the evidence

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

New model New model: Booster T2 (Booster Robotics) high confidence auto-approved (bulk)

2026-08-23T01:04:42+00:00

status commercial autonomy task-autonomous DoF 31 payload 10 kg deployments 3

Flagship announced July 2026; 140cm, 31 DoF base, up to 73 DoF with BrainCo Revo 3 hands; highest onboard compute in its class (2,000 TOPS).

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Booster T2
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

New model New model: Optimus Gen 3 (Tesla) medium confidence auto-approved (bulk)

2026-08-23T01:04:41+00:00

status pilot autonomy supervised-autonomy DoF 72 payload 20 kg deployments 4

Optimus Gen 3 enters pilot production in 2026 with 22-DoF hands, AI5 chip, and Grok integration. It is being tested in 24/7 autonomous shifts at Tesla factories.

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Optimus Gen 3
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim 'supervised-autonomy' not contradicted by the cited evidence
! Second-model check checked independently by glm: 1 claim(s) supported, 5 not addressed by the evidence

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

New model New model: Optimus Gen 4 (Tesla) low confidence auto-approved (bulk)

2026-08-23T01:04:41+00:00

status prototype autonomy research DoF — payload — kg

Optimus Gen 4 is an in-development iteration planned for high-volume manufacturing at Gigafactory Texas, with a focus on vertical integration and supply-chain redesign.

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (google.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Optimus Gen 4
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

New model New model: Go2 (Unitree Robotics) medium confidence auto-approved (bulk)

2026-08-23T01:04:40+00:00

status commercial autonomy supervised-autonomy DoF — payload — kg

The Go2 is a consumer/research quadruped robot, widely used in research and education, with extensive open-source support.

85.0/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): github.com, unitree.com
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 1 of 2 fetched page(s) mention Go2
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim 'supervised-autonomy' not contradicted by the cited evidence
! Second-model check checked independently by glm: 2 claim(s) supported, 1 not addressed by the evidence

2 sources: www.unitree.com github.com View robot →

New model New model: Go1 (Unitree Robotics) medium confidence auto-approved (bulk)

2026-08-23T01:04:40+00:00

status commercial autonomy supervised-autonomy DoF — payload — kg

The Go1 is a consumer quadruped robot, part of Unitree's open-source lineup, widely used in education and research.

85.0/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): github.com, unitree.com
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 1 of 2 fetched page(s) mention Go1
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim 'supervised-autonomy' not contradicted by the cited evidence
! Second-model check checked independently by glm: 2 claim(s) supported, 1 not addressed by the evidence

2 sources: www.unitree.com github.com View robot →

New model New model: Phoenix Gen 7 (Sanctuary AI) medium confidence auto-approved (bulk)

2026-08-23T01:04:39+00:00

status retired autonomy teleoperated DoF 40 payload 25 kg deployments 2

Gen 7 was the last bipedal Phoenix, claimed to learn tasks in under 24 hours (unverified). It was retired from commercial marketing in June 2026 when Sanctuary pivoted to Physical AI software.

85.0/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
✓ Source independence 2 independent domain(s): betakit.com, theagenttimes.com
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Phoenix Gen 7
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim no elevated autonomy claim to challenge
! Second-model check checked independently by glm: 1 claim(s) supported, 5 not addressed by the evidence

2 sources: theagenttimes.com betakit.com View robot →

New model New model: Unitree H1-2 (Unitree Robotics) medium confidence auto-approved (bulk)

2026-08-23T01:04:39+00:00

status commercial autonomy supervised-autonomy DoF — payload — kg

Unitree H1-2 is an updated version of the H1 humanoid, listed on Unitree's official website.

77.5/100 validated with warnings · 2/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (unitree.com) — not independent corroboration
✓ Citations resolve 2 of 2 cited URLs resolved
✓ Citations on-topic 2 of 2 fetched page(s) mention Unitree H1-2
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim 'supervised-autonomy' not contradicted by the cited evidence
! Second-model check checked independently by glm: 1 claim(s) supported, 2 not addressed by the evidence

2 sources: unitree.com unitree.com View robot →

New model New model: Moxi (Diligent Robotics) high confidence auto-approved (bulk)

2026-08-23T01:04:39+00:00

status commercial autonomy task-autonomous DoF — payload — kg deployments 3

Moxi is a wheeled humanoid robot designed for hospital logistics, autonomously delivering supplies and lab samples. It has been deployed at multiple named hospital systems including Texas Health Resources and Cedars-Sinai.

62.5/100 validated with warnings · 1/2 citations resolved

✓ Schema record matches the schema
✓ Source count 2 source(s) cited
! Source independence all citations are on one domain (diligentrobots.com) — not independent corroboration
! Citations resolve 1 of 2 cited URLs did not resolve
✓ Citations on-topic 1 of 1 fetched page(s) mention Moxi
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude new record — no prior value to compare against
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by glm: 1 claim(s) supported, 3 not addressed by the evidence

2 sources: www.diligentrobots.com www.diligentrobots.com View robot →

Spec update Update HRP-5P high confidence human-approved

2026-08-23T00:51:41+00:00

Field	Before		After
autonomy_level	supervised-autonomy	→	task-autonomous
ai_model	Object recognition and environment-measurement software for autonomous heavy-labor task execution	→	No VLM on-board; no foundation models; no imitation learning. Intelligence stack comprises: 3D environmental measurement, AI-based object detection, full-body motion planning/control, task description + execution management
onboard_compute	—	→	Custom software stack; specific hardware not publicly disclosed
payload_kg	—	→	13
runtime_hours	—	→	1
hand_dof_per_hand	—	→	2
dexterous_hands	True	→	False

85.0/100 validated with warnings · 6/6 citations resolved

✓ Schema record matches the schema
✓ Source count 12 source(s) cited
✓ Source independence 6 independent domain(s): aijack.info, firgellirobots.com, humanoid.press, robolist.ai, techacute.com, thehumanoidrace.com
✓ Citations resolve 6 of 6 cited URLs resolved
✓ Citations on-topic 5 of 6 fetched page(s) mention HRP-5P
✓ Value plausibility all numeric values within plausible ranges
✓ Internal consistency record is internally consistent
✓ Change magnitude changes are proportionate to the current values
✓ Autonomy claim 'task-autonomous' not contradicted by the cited evidence
! Second-model check checked independently by glm: the cited evidence does not address any of the claims

12 sources: www.robolist.ai www.humanoid.press techacute.com aijack.info View robot →

New model New model: Optimus Gen 3 (Tesla) medium confidence human-approved

2026-07-17T03:45:47+00:00

status prototype autonomy task-autonomous DoF 40 payload 20 kg deployments 1

Production-intent model with AI5 compute, tendon-driven hands, and 40+ DoF; production slated for late July/August 2026.

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

New model New model: Optimus Gen 2.5 (Tesla) medium confidence human-approved

2026-07-17T03:45:45+00:00

status prototype autonomy supervised-autonomy DoF 28 payload 20 kg deployments 2

Transitional model between Gen 2 and Gen 3, introduced mid-2025 with quieter brushless motors and upgraded 22-DoF hands.

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update Figure 02 high confidence human-approved

2026-07-17T03:45:43+00:00

Field	Before		After
status	pilot	→	retired
autonomy_level	task-autonomous	→	supervised-autonomy
ai_model	Helix (in-house VLA, dual-system 7-9 Hz + 200 Hz)	→	Helix VLA / OpenAI speech and vision-language models
onboard_compute	Dual NVIDIA RTX-based GPU modules (~3x Figure 01 inference)	→	Dual NVIDIA RTX GPU-based modules
deployments	BMW Group Plant Spartanburg (10-month pilot, ~1,250 hrs across 30,000+ X3 vehicles)	→	BMW Group Plant Spartanburg (completed 11-month pilot in November 2025, ~1,250 hrs across 30,000+ X3 vehicles, handling 90,000+ parts with >99% accuracy)

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update Figure 03 high confidence human-approved

2026-07-17T03:45:41+00:00

Field	Before		After
status	prototype	→	pilot
ai_model	Helix (vision-language-action model)	→	Helix 02
onboard_compute	dual low-power embedded GPUs	→	Dual low-power embedded GPUs
weight_kg	60	→	61
total_dof	—	→	44
hand_dof_per_hand	—	→	16
price_usd	—	→	150000
deployments		→	BMW Spartanburg plant (deployed June 2026 for production logistics/sequencing), Brookfield Asset Management (Project Go-Big strategic partnership, September 2025), Catalyst Brands (logistics scaling agreement, May 2026)

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update Figure 01 high confidence human-approved

2026-07-17T03:45:39+00:00

Field	Before		After
status	prototype	→	retired
ai_model	OpenAI vision-language model (speech-to-speech integration)	→	Helix / OpenAI VLM
onboard_compute	—	→	High-performance onboard computer
hand_dof_per_hand	—	→	6
price_usd	—	→	150000
funding_usd_millions	—	→	2675
valuation_usd_billions	—	→	39
deployments		→	BMW Spartanburg (initial testing/agreement in January 2024)

2 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Spec update Update Atlas (electric) high confidence human-approved

2026-07-17T03:45:36+00:00

Field	Before		After
status	limited-production	→	commercial
autonomy_level	supervised-autonomy	→	task-autonomous
ai_model	NVIDIA Isaac GR00T + Google DeepMind Gemini Robotics; Toyota Research Institute LBM	→	Google DeepMind Gemini Robotics (Gemini 3) + Toyota Research Institute Large Behavior Models + Boston Dynamics Orbit fleet software
onboard_compute	NVIDIA Jetson Thor	→	NVIDIA Jetson Thor (Blackwell, 2070 FP4 TFLOPS)
hand_dof_per_hand	—	→	7
deployments	Hyundai HMGMA (Georgia metaplant, automotive assembly), Google DeepMind (AI research)	→	Hyundai HMGMA (Georgia metaplant, automotive assembly), Google DeepMind (AI research), Kia Motors (U.S. manufacturing plants), FIFA World Cup 2026 (ceremonial ball delivery at MetLife Stadium)

6 sources: vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com vertexaisearch.cloud.google.com View robot →

Company update Update Boston Dynamics human-approved

2026-07-17T03:45:31+00:00