FrontierSWE V2 Best@5
Code
FrontierSWE V2 mean of the best normalized reward per task across up to five graded trials. This is not a confidence-interval bound.
Top models (higher is better)
| Model | Score |
|---|---|
| GPT-6 Astra | 73.0 |
| Claude Opus 5.5 | 71.5 |
| Claude Opus 5 | 60.7 |
| GPT-5.6 Sol | 44.2 |
| Grok 4.7 | 41.1 |
| GLM-5.3 | 40.6 |
| Kimi K3 | 37.6 |
| Grok 4.6 | 37.3 |
| Gemini 3.8 Flash | 31.6 |
| Gemini 3.7 Flash | 31.2 |
| Muse Spark 1.2 | 18.4 |
| Inkling | 10.8 |