RoboDojo: Unified Benchmark for Sim and Real-World Robots
量子位 · wechat · 2026-07-08
The RoboTwin team introduced RoboDojo, a new robotic manipulation benchmark featuring 42 simulated tasks and 18 real-world tasks. It evaluates five core capabilities: generalization, memory, precise manipulation, long-horizon execution, and open-vocabulary semantic understanding. The benchmark highly standardizes real-world evaluations through unified hardware setups, scene resets, and a double-blind scoring mechanism. Under these rigorous tests, the current state-of-the-art generalist robot policy achieved an average success rate of only 8.80% in simulation and a maximum of 12.8% in the real world, compared to a 100% success rate by human experts. This highlights a significant gap between current robotic foundation models and true general physical manipulation.
Related event: RoboDojo: Unified Sim-and-Real Robotics Benchmark(2 posts)→
More from Embodied
- Creator buys a Reachy robot and asks what to build first — dee_hw · 2026-07-21
- MW team shows Gen1 of MW-bot, a semi-humanoid home robot built for pantry storage and ceiling rails — CyberRobooo · 2026-07-21
- Kunlun Spins Out Riemann Dynamics, Unveils World Action Model for Robotics — 0xAllen_ · 2026-07-21
- Gritt exits stealth with $34M to automate construction sites, starting with solar plants — TechCrunch AI · 2026-07-21
- BrainCo demos near-real-time bionics without implants and claims 85% lower prosthetic cost — TrueOrange9944 · 2026-07-21
- OpenAI’s $230 CodexMicro sold out, and users are already cloning it with Stream Decks — APPSO · 2026-07-21