AtmosCoder-Bench: leaderboard
Metric: Accuracy (%, mean of 3 runs). Source: github.com. 16 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.5 | 97.6 |
| 2 | Gemini 3.1 Pro (Preview) (Thinking) | 96 |
| 3 | Kimi K2.6 (Thinking) | 93.8 |
| 4 | DeepSeek V4 Pro (Reasoning) | 93.3 |
| 5 | DeepSeek V4 Flash (Reasoning) | 91.1 |
| 6 | GPT-5.5 (Non-reasoning) | 90.8 |
| 7 | Kimi K2.6 (Non-reasoning) | 90.1 |
| 8 | Qwen 3.6 27B (Reasoning) | 88.5 |
| 9 | DeepSeek V4 Pro (Non-reasoning) | 82.6 |
| 10 | DeepSeek V4 Flash (Non-reasoning) | 81.7 |
| 11 | Qwen 3.6 27B (Non-reasoning) | 79.4 |
| 12 | Qwen 3.5 9B (Reasoning) | 76.9 |
| 13 | Qwen 3.5 9B (Non-reasoning) | 63.4 |
Interactive version: theaggregate.ai/benchmark?slug=atmoscoder-bench · How It Works · Data refreshed daily, snapshot 2026-09-05.