Multi-turn ambiguous query clarification on ClarifyMT-Bench Contradictory persona 1.0
97.4AccuracyGemini-2.5-Flash
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini-2.5-Flashvariant=Flash2025.12 | 97.4 | |
| Claude-Sonnet-4.5variant=Sonnet2025.12 | 91.9 | |
| Llama-3.1-70B-Itparameters=70B, alignment=Instruction-Tuned2025.12 | 88.7 | |
| Llama-3.1-8B-Itparameters=8B, alignment=Instruction-Tuned2025.12 | 86.4 | |
| DeepSeek-V32025.12 | 85.6 | |
| GPT-4.12025.12 | 82 | |
| DeepSeek-R12025.12 | 73.7 | |
| Qwen-2.5-72B-Itparameters=72B, alignment=Instruction-Tuned2025.12 | 72.5 | |
| o32025.12 | 68.1 | |
| Qwen-2.5-7B-Itparameters=7B, alignment=Instruction-Tuned2025.12 | 42.4 |