Conversational Quality Evaluation on SKG-PROBE
0.766SKG-Eval ScoreGPTOSS-20B
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| GPTOSS-20Bnumber of parameters=20B2026.05 | 0.766 | 0.988 | 1 | 1 | — | |
| GEMMA-4-31Bnumber of parameters=31B2026.05 | 0.756 | 0.954 | 2 | 2 | — | |
| MINIMAX-M2.7number of parameters=M2.72026.05 | 0.742 | 0.971 | 3 | 3 | — | |
| LLAMA-3-70Bnumber of parameters=70B2026.05 | 0.741 | 0.994 | 4 | 4 | — | |
| DEEPSEEK-V4-PRO2026.05 | 0.734 | 0.962 | 5 | 5 | — | |
| MISTRAL-7Bnumber of parameters=7B2026.05 | 0.709 | 0.951 | 6 | 6 | — | |
| LLM-as-Judge2026.05 | — | — | — | — | 0.73 | |
| SKG-Eval2026.05 | — | — | — | — | 1 |