SWE Atlas: leaderboard
Metric: Score (self-reported). Source: benchmarklist.com. 38 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 5 | 63.2 |
| 2 | Qwen 3.8 Max | 60.3 |
| 3 | GLM-5.2 | 48.12 |
| 4 | GPT-5.6 Sol | 46 |
| 5 | GPT-5.6 Sol (xHigh) | 46 |
| 6 | GPT-5.5 (xHigh) | 45.43 |
| 7 | GPT-5.5 | 45.4 |
| 8 | GPT-5.4 (xHigh) | 40.8 |
| 9 | Claude Fable 5 | 39 |
| 10 | GPT-5.3 Codex | 32.6 |
| 11 | DeepSeek V4 Pro | 27.15 |
| 12 | Muse Spark | 24.2 |
| 13 | GLM-5 | 20.5 |
| 14 | Gemini 3.1 Pro (Preview) | 13.5 |
| 15 | Kimi K2.5 | 13.1 |
Interactive version: theaggregate.ai/benchmark?slug=swe-atlas · How It Works · Data refreshed daily, snapshot 2026-09-05.