FrontierPhysics: leaderboard
FrontierPhysics (You and colleagues, 2026; arXiv:2611.06910v1) measures end-to-end physics research on 56 physicist-authored tasks across nine domains. Agents review literature, plan and run research, then submit numerical outputs and a scientific manuscript. This board reports Score: Gated rubric score (0–100), mean over all 56 tasks and three trials per task. There are three trials per task, 168 designated rollouts per configuration and a 10-hour cap per rollout. Missing outcomes remain failures/zero scores. Each row preserves its native Codex, Claude Code, Grok Build or Antigravity harness and inference effort; exact harness versions and deployment checkpoints are not disclosed. All results use task-set v0.1.
Metric: Gated rubric score (0–100), mean over all 56 tasks and three trials per task; all verifiers and blocking rubrics must pass before non-blocking rubric quality contributes; v0.1, 10-hour cap per rollout; missing outcomes count as zero/failure; higher is better. Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 26 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5.5 (FrontierPhysics Run: v0.1; harness=Claude Code; effort=max; timeout=10h) | 26 | |
| 2 | GPT-6 Astra (FrontierPhysics Run: v0.1; harness=Codex; effort=max; timeout=10h) | 21.8 | |
| 3 | GPT-6.1 Sol (FrontierPhysics Run: v0.1; harness=Codex; effort=max; timeout=10h) | 18.4 | |
| 4 | GPT-6 Astra (FrontierPhysics Run: v0.1; harness=Codex; effort=xhigh; timeout=10h) | 18 | |
| 5 | GPT-6.1 Terra (FrontierPhysics Run: v0.1; harness=Codex; effort=max; timeout=10h) | 12.6 | |
| 6 | GPT-6 Sol (FrontierPhysics Run: v0.1; harness=Codex; effort=max; timeout=10h) | 12.5 | |
| 7 | Claude Fable 5.1 (FrontierPhysics Run: v0.1; harness=Claude Code; effort=max; timeout=10h) | 10.2 | |
| 8 | Claude Opus 5 (FrontierPhysics Run: v0.1; harness=Claude Code; effort=max; timeout=10h) | 8.6 | |
| 9 | Gemini 4 Argon (FrontierPhysics Run: v0.1; harness=Antigravity; effort=high; timeout=10h) | 6.5 | |
| 10 | GPT-5.6 Sol (FrontierPhysics Run: v0.1; harness=Codex; effort=max; timeout=10h) | 6.2 |
Interactive version: theaggregate.ai/benchmark?slug=frontierphysics · How It Works · Data refreshed daily, snapshot 2026-11-06.