Faraday Future and its robotics arm FFR say their embodied AI system, OmniMind, has taken the top spot on a global leaderboard that tests robot intelligence in simulation. The companies announced the result in a weekly business update published on 9 November 2026, saying the model ranked first overall on RoboColiseum as of 6 November.

OmniMind scored 0.8698 out of 1, according to the update from founder and global chief executive YT Jia. That put it ahead of NVIDIA GR00T N1.7, GPT-6 Astra, Claude Opus 5.5 and π0.5, the rival models named in the announcement. RoboColiseum tests embodied AI models in simulated environments rather than on physical robots, and ranks them on a public leaderboard.

What the score is made of

The announcement broke the headline figure into three parts. OmniMind scored about 0.99 in instruction following, 0.92 in spatial reasoning and 0.96 in robustness to disturbances. Jia described OmniMind as a layered system rather than a single action model: it interprets instructions, uses 3D perception to locate a target, and then generates a path through motion planning.

The companies presented the ranking as evidence of progress in what they call the EAI Brain, the core of their embodied AI work. Jia said the result completes Phase One of the company’s “Built in USA” programme. The update also said FFR is advancing sales with several leading US integrators of security and inspection robotics, although it gave no sales figures, revenue or deployment numbers.

What the ranking does not show

The result is a simulation benchmark, and the companies were clear about that in the announcement itself. A high score in RoboColiseum does not show that OmniMind can do paid work on real robots in the field, and the update gave no figures to show commercial impact. The ranking is also company-reported, drawn from the leaderboard as of 6 November, and leaderboards can change as new models are tested.

Faraday Future, listed on the Nasdaq as FFAI, has been shifting its focus towards embodied AI alongside its electric vehicle business. The weekly update was the 80th in the series Jia has published for the company.

Earnings dates set for November

The same update set dates for the companies’ next earnings calls. FFR will hold its third-quarter 2026 earnings call on 12 November at 4 p.m. Pacific Time, its first since its listing. FFAI will hold its third-quarter 2026 earnings call on 16 November at 4 p.m. Pacific Time.

Investors and robotics watchers will be looking at those calls for the detail the weekly update did not give: whether the simulation lead is turning into orders, deployments or revenue, and when OmniMind will be tested on physical robots rather than in simulation.