Embodied First Exam: 1,090 robotics tasks in one line of code, backed by $500K Robocurve fund
ManlingLi_ · x · 2026-10-20
Manling Li and colleagues launched Embodied First Exam, an open-source embodied benchmark that runs coding agents on 1,090 tasks across 31 simulation suites and 44 robots with one line of code — covering puzzles with spatial constraints, multi-robot collaboration, real-time eval where the world won't pause, agile whole-body control, and sim2sim AutoResearch.
The companion Robocurve funding program offers:
- $500K total pool: $20K cash per team, $2K compute credits, $1K in evaluation materials
- A pair of YAM arms (ABC Research Kit with 3× Intel RealSense D405 cameras) that teams keep afterward
- Requirements: design a bimanual benchmark with at least 20 tasks and run real-world evals across 3+ publicly accessible models (VLAs like MolmoAct and Isaac 0.5, LLMs like Gemini 3.7 Flash)
- Output: co-authorship on a mega-author paper
Supporters note that measuring robotics capability over time is a crucial input for forecasting AI takeoff, yet widely-used high-quality robotics benchmarks are essentially nonexistent.
More from Embodied
- Robotics take: IK guarantees no precision edge, hardware—not the solver—is the bottleneck — ChongZzZhang · 2026-10-20
- MPC vs RL for robot tracking: the precision bottleneck is hardware, not the solver — ChongZzZhang · 2026-10-20
- The most important AI device of the next decade won't be a smartphone — VraserX · 2026-10-20
- NUS MAGIC Lab is hiring a postdoc, 2 PhD students and 3 research assistants for robotics — DJiafei · 2026-10-20
- AlphaNIST two-finger gripper screws nuts multiple ways, "it's just PPO" — _Stocko_ · 2026-10-20
- New RL paper: C-Slerp learns 5-cube stacking from scratch with no rewards or pretraining — ben_eysenbach · 2026-10-20