Coding Agent Capacity on Scicode
57.7ScoreGemini-3-pro High
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini-3-pro High2026.06 | 57.7 | |
| Gemini-3-Flash High2026.06 | 55 | |
| Claude-Opus-4.52026.06 | 52.8 | |
| Seed2.0 Lite2026.06 | 52.4 | |
| GPT-5.2 High2026.06 | 49.7 | |
| Seed2.0 Pro2026.06 | 48.5 | |
| Claude-Sonnet-4.52026.06 | 47.9 | |
| GPT-5-mini High2026.06 | 40.2 |