CONCUR - Pass@3: leaderboard
Metric: Pass@3 (%): share of the 115 CONCUR problems for which at least one of three independently generated Java 8 programs compiles and passes Java PathFinder model checking; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 23 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 91.3 | #91 |
| 2 | GPT-4o | 82.61 | #333 |
| 3 | Claude Opus 4.1 | 79.13 | #132 |
| 4 | Gemini 3 Pro | 73.91 | #77 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=concur-pass-3 · How It Works · Data refreshed daily, snapshot 2026-10-11.