Three-month-old Simate tops RoboDojo benchmark with its first physical AI model
量子位 · wechat · 2026-09-26
Simate, a three-month-old startup founded by ex-autonomous-driving experts, released Simate-beta, a general-purpose "fast system" for physical AI that tops the RoboDojo leaderboard (avg score 33.95, SR 27.96%) with models not optimized for the benchmark.
- Architecture bets on large-parameter, edge-deployable fast systems with 4D physical perception and hierarchical temporal memory, aiming for zero-shot generalization alongside slow reasoning models.
- Its AI-native R&D loop (SiPAI + AutoResearch + Sinfra infra) lets agents run dozens of parallel research lines, filtering via world models and simulation before real-robot tests.
- Founders include Zhang Ying (led end-to-end autonomous driving comparable to Tesla FSD) and HKUST professor Zhan Fangneng.
- The company has raised multiple RMB hundreds-of-millions rounds and plans staged papers, open-source releases, and year-end results on zero-shot generalization.
More from Embodied
- Lighthouse extends Obsidian welding foundation model to run on any robotic platform — bennash · 2026-11-11
- Sipeed's MaixStation Dock PCB returns, 14-layer SOM still in fab — AIFlow_ML · 2026-11-11
- NAE EU-US Frontiers of Engineering symposium explores AI-driven materials, robotics, neuromorphic computing — ProfBuehlerMIT · 2026-11-11
- Paper link for FLEX: robots generalize language corrections into preferences — JafarNajafov · 2026-11-11
- MIT and NYU's FLEX lets robots generalize a single spoken correction into broad preferences — JafarNajafov · 2026-11-11
- AI lets robotics research thrive in labs with no robotics experts — 4310sy · 2026-11-11