Natural Language to Optimization on ComplexOR (test)
72.2Success Rate (SR)NC2C
Evaluation results
| Method | Links | ||
|---|---|---|---|
| NC2CModel=GPT-5.12026.01 | 72.2 | 80.6 | |
| NC2CModel=Qwen3-235B-A22B2026.01 | 66.7 | 75 | |
| OptiMUSModel=GPT-5.12026.01 | 63.9 | 72.2 | |
| Chain-of-ExpertsModel=GPT-5.12026.01 | 58.3 | 66.7 | |
| ReflexionModel=GPT-5.12026.01 | 55.6 | 64.4 | |
| OptiMUSModel=Qwen3-235B-A22B2026.01 | 55.6 | 63.9 | |
| VanillaModel=GPT-5.12026.01 | 52.8 | 60 | |
| Chain-of-ExpertsModel=Qwen3-235B-A22B2026.01 | 50 | 58.3 | |
| ReflexionModel=Qwen3-235B-A22B2026.01 | 48.3 | 56.1 | |
| VanillaModel=Qwen3-235B-A22B2026.01 | 45 | 52.8 |