PerfReasoning - Max Effort: leaderboard
Metric: Both-Exact Pass Rate (%). Source: arxiv.org. 4 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.6 Sol | 88.4 |
| 2 | GPT-5.6 Terra | 3.7 |
| 3 | DeepSeek V4 Pro | 3.2 |
| 4 | GPT-5.6 Luna | 0 |
Interactive version: theaggregate.ai/benchmark?slug=perfreasoning-max-effort · How It Works · Data refreshed daily, snapshot 2026-09-21.