Coding on OJBench
40.7AccuracyLongCat Flash
Evaluation Results
| Method | Links | |
|---|---|---|
| LongCat FlashParams=560B2026.06 | 40.7 | |
| VibeThinker-3BParams=3B2026.06 | 38.6 | |
| Qwen3-235B-A22B-ThinkingParams=235B2026.06 | 32.5 | |
| OpenReasoning-NemotronParams=7B2026.06 | 25.9 | |
| Qwen3.5-4BParams=4B2026.06 | 23.5 | |
| Mimo7B-RL-0530Params=7B2026.06 | 20.2 | |
| Qwen3-4B-Thinking-2507Params=4B2026.06 | 17.9 | |
| Olmo-3-ThinkParams=7B2026.06 | 15.6 | |
| Ministral-3-Reasoning-2512Params=14B2026.06 | 15.1 | |
| Phi4-Reasoning-PlusParams=14B2026.06 | 14.4 | |
| Hunyuan-4B-InstructParams=4B2026.06 | 12.1 | |
| SmolLM3Params=3B2026.06 | 5.2 |