Robotics
Task 37 / 47
DynamicObstacleNavigation
Navigate a differential-drive robot from start to goal among moving obstacles under velocity/acceleration limits while avoiding collisions and minimizing time or path length. Simplified dynamics/obstacle laws in simulation score collisions, success, and path cost—mobile robot motion planning engineering.
Model leaderboard
Updated v1 results · 2026-09-15 · raw score, higher is better.
| # | Participant | Raw score | Medal |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 0.086 | Gold |
| 2 | GPT-5.4 | 0.085714286 | Silver |
| 3 | GLM-5 | 0.0857 | Bronze |
| 4 | DeepSeek V3.2 | 0.0856 | — |
| 5 | Seed 2.0 Pro | 0.0855 | — |
| 6 | Gemini 3.1 Pro Preview | 0.0834 | — |
| 7 | Grok 4.20 | 0.0817 | — |
| 8 | Qwen3 Coder Next | 0.0765 | — |
Framework results · original paper
Historical results from the original evaluators · normalized score (0–100).
| # | Participant | Score |
|---|---|---|
| 1 | GPT-OSS + ABMCTS | 100.0 |
| 2 | GPT-OSS + ShinkaiEvolve | 99.3 |
| 3 | Claude Opus 4.6 + OpenEvolve | 98.6 |
| 4 | Claude Opus 4.6 + ShinkaiEvolve | 96.4 |
| 5 | Claude Opus 4.6 + ABMCTS | 93.5 |
| 6 | GPT-OSS + OpenEvolve | 0.0 |
Model results use the updated scores and medal thresholds. No valid score earns zero medal credit. Historical framework scores use the original evaluators and have not been updated alongside the model results.