Overall Performance on LiC Overall
76.4LiC ScoreQwen2.5-7B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen2.5-7BOptimization=RLAAR2025.10 | 76.4 | 67.6 | |
| Qwen3-8BOptimization=RLAAR2025.10 | 75.1 | 73.4 | |
| Qwen3-1.7BOptimization=RLAAR2025.10 | 71.9 | 68.4 | |
| Llama3.3-70B2025.10 | 69.6 | — | |
| GPT-4.12025.10 | 69.4 | 33.5 | |
| GPT-4o2025.10 | 65.2 | 31.5 | |
| Claude3-Haiku2025.10 | 65 | — | |
| Llama4-Scout2025.10 | 63.4 | — | |
| Qwen3-8BFinetuning=SFT, Distillation=Gemini-2.5-Pro2025.10 | 62.7 | 49 | |
| Qwen3-8Bsource=Grattafiori et al.2025.10 | 62.6 | 33.5 | |
| Gemini-2.5-Flash2025.10 | 62.5 | 33 | |
| GPT-o32025.10 | 62 | — | |
| Claude3.7-Sonnet2025.10 | 61 | — | |
| GPT4o-mini2025.10 | 60.8 | — | |
| Qwen3-1.7Bsource=Grattafiori et al.2025.10 | 58.8 | 23 | |
| DeepSeek-R12025.10 | 57.4 | — | |
| OLMo22025.10 | 57.3 | — | |
| Gemini-2.5-Pro2025.10 | 56.6 | 35.6 | |
| Phi-42025.10 | 56.4 | — | |
| Qwen2.5-7BFinetuning=SFT, Distillation=Gemini-2.5-Pro2025.10 | 55 | 50.1 | |
| Qwen2.5-7B2025.10 | 54.6 | 30.5 |