Path Planning Optimality Proofs (With Lemmas): leaderboard
Metric: Success rate (%): share of tasks with both the final-answer and the reasoning score at least 7 of 10, given the problem, the algorithm and the key lemmas of the source paper (Setting-2) on the benchmark's 34 research-level approximation-ratio proof tasks for robotic path-planning algorithms (collected from 11 papers, problem and algorithm names anonymized), reasoning mode on, no reasoning template; a GPT-5.2 judge grades each proof against the source paper's ground-truth proof; higher is better. Source: arxiv.org. Saturation forecast: Around March 2028. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 397B A17B (Thinking) | 44.12 | #141 (Qwen 3.5 397B A17B) |
| 2 | Gemini 3 Pro | 35.29 | #77 |
| 3 | GPT-5.2 (Thinking) | 32.35 | #105 (GPT-5.2) |
| 4 | Grok 4.1 (Thinking) | 29.41 | #218 (Grok 4.1) |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=path-planning-optimality-proofs-with-lemmas · How It Works · Data refreshed daily, snapshot 2026-10-11.