datacurve/
Run by mupt-ai · Released 2026-11-09 · 113 tasks · 7 settings
Model Tradeoffs
Which coding agent works best on datacurve/deep-swe-1-1? Most accurate: GPT-6.1 Sol (X-High), 75.2% at $0.83 per task. Best for less: GPT-6.1 Sol (High), 71.7% at $0.69. 7 model settings scored on 113 tasks from the benchmark.
All Settings
| Model | Harness | Reasoning | Access | Accuracy | Cost / Task | Frontier |
|---|---|---|---|---|---|---|
| GPT-6.1 Sol | Codex | X-High | API Key | 75.2% | $0.83 | Yes |
| GPT-6.1 Sol | Pi | High | API Key | 71.7% | $0.69 | Yes |
| shift-1 | Codex | Default | Custom Endpoint | 69.0% | $3.68 | No |
| switchyard | Codex | Default | Custom Endpoint | 68.1% | $0.71 | No |
| Dari Router | Pi | Default | API Key | 65.5% | $0.57 | Yes |
| accounts/fireworks/routers/kimi-fast-latest | Pi | Default | Custom Endpoint | 61.1% | $2.77 | No |
| GPT-6 Luna | Codex | High | API Key | 39.8% | $0.076 | Yes |