Robotics
Task 38 / 47
PIDTuning
Tune cascaded PID gains for a 2D quadrotor across multiple flight scenarios to reduce tracking error and overshoot under actuator saturation. Continuous gain vectors are scored by integrated error and stability-style constraints—low-level aerial vehicle controls engineering.
Model leaderboard
Updated v1 results · 2026-09-15 · raw score, higher is better.
| # | Participant | Raw score | Medal |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 0.1632 | Gold |
| 2 | Grok 4.20 | 0.1585 | Silver |
| 3 | Gemini 3.1 Pro Preview | 0.1521 | Bronze |
| 4 | GLM-5 | 0.1515 | — |
| 5 | Seed 2.0 Pro | 0.1514 | — |
| 6 | GPT-5.4 | 0.15111728 | — |
| 7 | DeepSeek V3.2 | 0.151 | — |
| 8 | Qwen3 Coder Next | 0.1422 | — |
Framework results · original paper
Historical results from the original evaluators · normalized score (0–100).
| # | Participant | Score |
|---|---|---|
| 1 | Claude Opus 4.6 + ShinkaiEvolve | 100.0 |
| 2 | GPT-OSS + ShinkaiEvolve | 98.4 |
| 3 | Claude Opus 4.6 + OpenEvolve | 96.2 |
| 4 | Claude Opus 4.6 + ABMCTS | 95.7 |
| 5 | GPT-OSS + ABMCTS | 0.5 |
| 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.