PhysRSI - Success: leaderboard
PhysRSI-eval (2026): six model/runtime systems on 18 robotic simulation tasks; initial K0 scores averaged over three runs per task, followed by iterative K1–K17 revisions. Codex uses 4200-second reasoning/1800-second revision budgets; Claude Code uses 4200/600. Runtime settings are part of these descriptive system comparisons. Normalized Score (0–1); higher is better.
Metric: Normalized Score (0–1). Source: arxiv.org. Saturation forecast: Around January 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6 Astra (xHigh) | 0.85 | #6 (GPT-6 Astra) |
| 2 | Claude Opus 5.5 (High) | 0.74 | #4 (Claude Opus 5.5) |
| 3 | GPT-6.1 Sol (xHigh) | 0.7 | #8 (GPT-6.1 Sol) |
| 4 | Claude Sonnet 5.5 (xHigh) | 0.7 | #11 (Claude Sonnet 5.5) |
| 5 | GPT-6 Sol (xHigh) | 0.47 | #16 (GPT-6 Sol) |
| 6 | GPT-5.6 Sol (xHigh) | 0.1 | #18 (GPT-5.6 Sol) |
No result here: #7 Gemini 4 Argon, #9 Claude Fable 5.1, #10 Muse Wonder Full.
Interactive version: theaggregate.ai/benchmark?slug=physrsi-success · How It Works · Data refreshed daily, snapshot 2026-10-21.