Sphinx - C#: leaderboard
Metric: Checklist Coverage Score (%; GPT-4o judge, 450 buggy + 50 bug-free PRs per language). Source: arxiv.org. Saturation forecast: Estimated already saturated. 29 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | O4 Mini | 40.01 |
| 2 | GPT-4.1 | 38.62 |
| 3 | Claude 3.7 Sonnet | 37.68 |
| 4 | Claude 3.5 Sonnet | 37.16 |
| 5 | Gemini 2.5 Pro | 36.75 |
| 6 | Qwen 2.5 72B Instruct | 35.06 |
| 7 | O3 Mini | 34.94 |
| 8 | GPT-4o (2024-08-06) | 34.32 |
| 9 | Qwen 3 235B A22B | 34.31 |
| 10 | GPT-4o Mini | 34.13 |
| 11 | Qwen 3 30B A3B | 33.03 |
| 12 | GPT-4.1 Mini | 32.38 |
| 13 | GPT-4.1 Nano | 32.14 |
| 14 | DeepSeek V3 | 31.39 |
| 15 | DeepSeek R1 | 30.09 |
Interactive version: theaggregate.ai/benchmark?slug=sphinx-c · How It Works · Data refreshed daily, snapshot 2026-09-25.