RoboDojo-RC
Robocurve, Inc.Maintainer ↗6 RoboDojo tasks adapted, Cap Pen, Pack and Pour Fruit, Sort Objects, Stack Bowls, Stand Up Bottles, Store In Safe.
As of November 17, 2026, Dream-1 leads with 59.3, followed by Claude Fable 5.2 (57.3) and Gemini 4 Carbon (54.5). Official leaderboard ↗
Leaderboard
- Tasks
- 6
- Attempts / task
- 20
- Median cost / task
- $1.77
- Median time / task
- 4 min 02 s
Overall vs speed
Published time per task attempt
Overall vs cost
Published cost per task attempt
Progress score 0 to 100, 20 attempts per task.
Dream-159.331.843.060.8$0.202est.4 min 02 sest.2
Claude Fable 5.257.345.534.573.0$5.934 min 29 s3
Gemini 4 Carbon54.546.033.074.2$2.1632 min 03 s4
Gemini 4 Argon48.241.336.072.2$2.3420 min 30 s5
GPT-6 Astra42.546.829.046.6$3.483 min 58 s6
GPT-6.1 Sol41.747.035.045.9$2.154 min 02 s7
Claude Opus 5.538.249.329.539.3$2.233 min 06 s8
Gemini 3.8 Flash36.247.313.542.5$0.6607 min 13 s9
GPT-6 Sol22.043.518.511.8$0.9493 min 44 s10
MolmoAct218.910.02.513.4$0.221est.4 min 02 sest.11
GPT-5.6 Sol14.936.512.55.1$1.773 min 39 s12
GR00T N1.79.510.00.009.3$0.221est.4 min 02 sest.Tasks
6 tasks. Hover a task to play its clip (6 of 6).
▶ Clip
▶ Clip
▶ Clip
▶ Clip
▶ Clip
▶ ClipMethodology
- Scoring
- Staged rubric
- Time cap / attempt
- —
- Control rate
- —
- Models in index
- 12 of 12
Scoring. An LLM judge applies written rubric levels to each trial video.
Who runs it. Independent evaluator (Robocurve).
Robots. I2RT YAM (two-arm).
Note. Tests coding agents and robot models on the same tasks. Tasks chosen after trying GPT-6 Astra.
Time per attempt. Published or measured for 9 of 12 models. Estimated otherwise: median measured time on this board (9 models).
Cost per attempt. Published for 9 of 12 models. Estimated otherwise: Dream Labs API, $3 per robot-hour × time; one H100 at $3.29 per hour (Lambda on-demand, Oct 2026) × time. GPU price source ↗