# Vals SWE-bench → SciCode

Calibration · https://benchgap.net/calibration/1598 · data retrieved 2026-10-29

SciCode is estimated from Vals SWE-bench with a Michaelis–Menten curve fitted on 8 models measured on both: y = 1.2863·x / (1.20621 + x), R² = 0.31, cross-validated error 4.5 pp. It is used for 2 estimates.

| Estimated model | Vals SWE-bench | SciCode | Source |
|---|---|---|---|
| [Composer 2.5](https://benchgap.net/model/composer-2-5.md) | 79.6% | 51.1% | estimated ± 4.5 pp, medium confidence |
| [Gemini 3.1 Flash-Lite](https://benchgap.net/model/gemini-3-1-flash-lite.md) | 62.8% | 44.0% | estimated ± 4.5 pp, low confidence |

## Fitted on

| Model | Vals SWE-bench | SciCode |
|---|---|---|
| [Gemini 3.5 Flash](https://benchgap.net/model/gemini-3-5-flash.md) | 78.8% | 53.1% |
| [Grok 4.3](https://benchgap.net/model/grok-4-3.md) | 71.4% | 47.3% |
| [Inkling-Small](https://benchgap.net/model/inkling-small.md) | 82.2% | 48.7% |
| [Kimi K2.6](https://benchgap.net/model/kimi-2-6.md) | 76.2% | 52.2% |
| [Ling 3.0 Flash](https://benchgap.net/model/ling-3-0-flash.md) | 65.2% | 41.2% |
| [Nemotron 3 Ultra](https://benchgap.net/model/nemotron-3-ultra.md) | 69.0% | 44.6% |
| [Qwen 3.6 Max (preview)](https://benchgap.net/model/qwen3-6-max-preview.md) | 72.8% | 47.0% |
| [Qwen3.7 Max](https://benchgap.net/model/qwen3-7-max.md) | 68.8% | 53.5% |