Photo via South China Morning Post

Maxwell smart: new Chinese embodied model tops global AI ranking on physical tasks

Read the original at South China Morning Post ↗
Cite

Tong, Z. (2026, October 1). Maxwell smart: new Chinese embodied model tops global AI ranking on physical tasks. South China Morning Post. https://www.scmp.com/news/china/science/article/3369420/maxwell-smart-new-chinese-embodied-model-tops-global-ai-ranking-physical-tasks?utm_source=rss_feed

Dated when The Aporia collected the article, usually within hours of publication. Check the original for its exact date and byline.

South China Morning Post · collected 2026-10-01 · by Zhang Tong

Quick Summary

A new Chinese AI model called Maxwell has topped the global ranking on physical tasks, scoring 91.9 on the Meta-World benchmark, which is the highest recorded score ever. Developed by the CAS Institute of Artificial Intelligence for Industries, Maxwell excels in performing 50 everyday tasks such as grasping objects and opening doors. This list includes submissions from numerous international teams like Google DeepMind and MIT, with the second-place entry scoring 90 and the third-place entry scoring 88.3.
Written by AI on 2026-10-01, from this article's own text only (the story summary below draws on all the coverage).

AI analysis

Story summary

Developed by the CAS Institute of Artificial Intelligence for Industries, the robot model Maxwell has achieved a top score of 91.9 on the Meta-World benchmark, surpassing all previous entries in evaluating how well robots perform 50 everyday physical tasks such as grasping and carrying objects, opening doors, and using drawers. Established by researchers from institutions like Stanford University and UC Berkeley, this benchmark assesses global teams' capabilities, including those from Google DeepMind, Alibaba, Meituan, Carnegie Mellon University, MIT, the University of Cambridge, and the Toyota Research Institute. In recent submissions as of July, FabriVLA (developed by Shenzhen-based Youibot) came in second with a score of 90, while SUREFlow, developed by researchers at Kyungpook National University in South Korea, placed third with 88.3.

Written for “Chinese AI Model Leads Ranking” on 2026-11-02, grounded in this article and the 0 other(s) covering the same event.

Story

📰 Chinese AI Model Leads Ranking
Technology · 1 article(s) covering the same event.

Compared with similar articles

Nothing to compare against. No article is close enough to this one for the pipeline to have linked or judged the pair.
How this article was measured signals, leaning charts, publisher record, byline, topics and claims

Signals How these are calculated →

Claims extracted
6
sentences that state something as happening or true
Uncertain
0%
0 of 6 hedged ("reportedly", "may", "according to")
Leaning
not political
takes no side on a contested political question
Publisher signals
63/100
South China Morning Post's corrections and hedging, across everything collected from it. Not a measure of accuracy.
Outlets on this story
1
Technology
Same framing elsewhere
1
articles that frame the event the way this one does
Analyzed 2026-10-01 · how these are computed

How this is being covered How these are calculated →

Article leaning vs. publisher reliability
Source leaning vs. consistency

Publisher

South China Morning Post · 3608 article(s) · 7 correction(s) detected
Running correction rate · 7 correction(s)
2026-11-01
Former Apple Daily publisher joins mainland China tour for rehabilitated offenders
2026-10-16
Nobel Peace Prize: the message behind Navi Pillay’s win
2026-10-16
Indonesian governor’s Singapore trip adds fury to forest fire, haze crisis
2026-10-04
Flight to Okinawa diverts back to Hong Kong after suspected landing gear fault
2026-10-04
Tennessee prisons chief resigning after botched execution of Christa Pike
2026-09-29
Indonesian prison chief suspended over secret VIP compound for inmates
2026-09-28
Trump ‘arms offer’ to China jolts Washington into moves to contain the fallout

Who wrote this

Zhang Tong
16 article(s) here · 0 carrying a prediction
🔮 “The programme noted in the letter will officially terminate after the current academic year,” she said.
2026-10-21 · mixed framing · US colleges scale back cooperation with Chinese institutions after Pentagon’s push
🔮 A federal judge subsequently acquitted him on seven additional counts of economic espionage, finding insufficient evidence that he knew or intended at the time of the thefts that his conduct would benefit the Chinese government.
2026-10-20 · assertive framing · US sanctions 5 Chinese entities over ex-Google engineer’s AI trade secrets theft
🔮 China lights up ‘super microscope’ hopes with world-leading electron gun beam By observing matter at the atomic scale, the Hefei Advanced Light Facility could drive breakthroughs in microchips
2026-09-28 · mixed framing · China lights up ‘super microscope’ hopes with world-leading electron gun beam
🔮 State broadcaster CCTV reported that the technology could fundamentally strengthen China’s ability to secure its own iron resources and reduce its heavy dependence on imports of the strategic commodity.
2026-09-06 · mixed framing · China begins mining iron from tailings, low-grade ore that can meet its demand for decades
🔮 The Fengyun-4C’s storm-tracking abilities could provide vital early warnings about extreme weather events during the rescue operation China is using the world’s most powerful geostationary weather satellite to track severe thunderstorms over Nepal and deliver real-time intelligence as the two countries continue search and rescue operations following last month’s devastating glacier collapse.
2026-09-06 · mixed framing · World’s most powerful lightning-chasing satellite joins China-Nepal flood relief efforts
🔮 By tracking the needle’s tiny deflections, the team can measure magnetic fields down to the femtotesla level – an advance that could widen the hunt for elusive dark matter particles or monitor brain signals.
2026-09-05 · assertive framing · Chinese-German team builds ultra-sensitive ‘floating compass’ to hunt dark matter
Also by Zhang Tong
US colleges scale back cooperation with Chinese institutions after Pentagon’s push
2026-10-21 · South China Morning Post
China close to matching US spending on ‘new way of war’, Republican senator warns
2026-10-20 · South China Morning Post
US sanctions 5 Chinese entities over ex-Google engineer’s AI trade secrets theft
2026-10-19 · South China Morning Post
New details emerge in US investigations of Chinese-American scientists over China links
2026-10-16 · South China Morning Post
Nothing else under this byline is closely related to this article, so these are simply their most recent.
All 16 articles by Zhang Tong →

Topics

Chinese Maxwell Meta Stanford University the CAS Institute of Artificial Intelligence for Industries

Subjects

Maxwell PERSON · 2× Alibaba ORG · 1× Carnegie Mellon University ORG · 1× Chinese NORP · 1× Google DeepMind ORG · 1× Meituan ORG · 1× Physical Intelligence ORG · 1× Stanford University ORG · 1× UC Berkeley ORG · 1× the CAS Institute of Artificial Intelligence for Industries ORG · 1×

Narrative

Developed by the CAS Institute of Artificial Intelligence for Industries, it scored 91.9 on the Meta-World benchmark – the highest recorded Maxwell scored 91.9 – the highest ever recorded on the simulation benchmark. Meta-World was set up by researchers from Stanford University, UC Berkeley and other institutions to evaluate how well robots perform 50 everyday physical tasks.
framing: assertive · carried by 1 article(s) · first seen 2026-10-01
2026-10-01 · South China Morning Post
Maxwell smart: new Chinese embodied model tops global AI ranking on physical tasks · assertive framing

Claims (6 extracted, 0 hedged)

Maxwell smart: new Chinese embodied model tops global AI ranking on physical tasks asserted
model → embody → tasks
Developed by the CAS Institute of Artificial Intelligence for Industries, it scored 91.9 on the Meta-World benchmark – the highest recorded Maxwell scored 91.9 – the highest ever recorded on the simulation benchmark. Meta-World was set up by researchers from Stanford University, UC Berkeley and other institutions to evaluate how well robots perform 50 everyday physical tasks. asserted
robots → develop → tasks
They range from grasping and carrying to opening doors and using drawers. asserted
They → range → drawers
The list draws submissions from leading teams worldwide, including Physical Intelligence, Google DeepMind, Alibaba, Meituan, Carnegie Mellon University, MIT, the University of Cambridge and the Toyota Research Institute. asserted
list → draw → Cambridge
Of the latest submissions, from July, second-placed FabriVLA – developed by Shenzhen-based Youibot – scored 90. asserted
FabriVLA → place → 90
SUREFlow, developed by researchers at Kyungpook National University in South Korea, was third with a score of 88.3. asserted
SUREFlow → develop → 88.3
💬Give feedback
🕘History 🎫Support
Terms · Privacy · beta v2.0.0
Articles to 2 Nov 2026