Tool-use Factuality Evaluation on InvivoQuery
78Factuality ScoreQwen3-30B-A3B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen3-30B-A3BExperimental Setup=Optimized2026.05 | 78 | 239 | |
| Mistral3.1-24B-ITExperimental Setup=Optimized2026.05 | 72 | 238 | |
| GPT-OSS-120BExperimental Setup=Optimized2026.05 | 71 | 235 | |
| Qwen-3-30B-ITExperimental Setup=Optimized2026.05 | 71 | 313 | |
| Gemma3-27B-ITExperimental Setup=Optimized2026.05 | 67 | 259 | |
| Qwen3-30B-A3BExperimental Setup=Always Tool2026.05 | 65 | 500 | |
| Qwen3-30B-A3BExperimental Setup=LNE2026.05 | 63 | 371 | |
| Qwen-3-30B-ITExperimental Setup=Always Tool2026.05 | 62 | 500 | |
| Qwen-3-30B-ITExperimental Setup=LNE2026.05 | 61 | 465 | |
| Llama3.2-3B-ITExperimental Setup=Optimized2026.05 | 61 | 251 | |
| Qwen3-30B-A3BExperimental Setup=LUEx2026.05 | 60 | 176 | |
| GPT-OSS-120BExperimental Setup=Always Tool2026.05 | 59 | 500 | |
| GPT-OSS-120BExperimental Setup=LNE2026.05 | 59 | 442 | |
| Qwen3-30B-A3BExperimental Setup=LUEx,d2026.05 | 59 | 177 | |
| Gemma3-27B-ITExperimental Setup=Always Tool2026.05 | 59 | 500 | |
| Mistral3.1-24B-ITExperimental Setup=LNE2026.05 | 59 | 478 | |
| Qwen3-30B-A3BExperimental Setup=No Tool2026.05 | 58 | 0 | |
| Gemma3-27B-ITExperimental Setup=LNE2026.05 | 58 | 469 | |
| Mistral3.1-24B-ITExperimental Setup=Always Tool2026.05 | 58 | 500 | |
| GPT-OSS-120BExperimental Setup=LUEx,d2026.05 | 57 | 163 | |
| GPT-OSS-120BExperimental Setup=Self-decided2026.05 | 56 | 433 | |
| GPT-OSS-120BExperimental Setup=LUEx2026.05 | 56 | 164 | |
| Gemma3-27B-ITExperimental Setup=LUEx2026.05 | 56 | 358 | |
| Gemma3-27B-ITExperimental Setup=LUEx,d2026.05 | 56 | 344 | |
| Qwen-3-30B-ITExperimental Setup=LUEx2026.05 | 55 | 269 | |
| Mistral3.1-24B-ITExperimental Setup=No Tool2026.05 | 55 | 0 | |
| Mistral3.1-24B-ITExperimental Setup=LUEx,d2026.05 | 55 | 93 | |
| Qwen-3-30B-ITExperimental Setup=LUEx,d2026.05 | 54 | 250 | |
| Mistral3.1-24B-ITExperimental Setup=Self-decided2026.05 | 54 | 270 | |
| Mistral3.1-24B-ITExperimental Setup=LUEx2026.05 | 54 | 112 | |
| GPT-OSS-120BExperimental Setup=No Tool2026.05 | 53 | 0 | |
| Gemma3-27B-ITExperimental Setup=No Tool2026.05 | 50 | 0 | |
| Llama3.2-3B-ITExperimental Setup=Always Tool2026.05 | 50 | 500 | |
| Llama3.2-3B-ITExperimental Setup=LNE2026.05 | 50 | 465 | |
| Llama3.2-3B-ITExperimental Setup=LUEx2026.05 | 48 | 237 | |
| Llama3.2-3B-ITExperimental Setup=LUEx,d2026.05 | 48 | 211 | |
| Qwen-3-30B-ITExperimental Setup=No Tool2026.05 | 46 | 0 | |
| Llama3.2-3B-ITExperimental Setup=No Tool2026.05 | 45 | 0 | |
| Llama3.2-3B-ITExperimental Setup=Self-decided2026.05 | 44 | 1,376 | |
| Gemma3-27B-ITExperimental Setup=Self-decided2026.05 | 42 | 326 | |
| Qwen3-30B-A3BExperimental Setup=Self-decided2026.05 | 41 | 235 | |
| Qwen-3-30B-ITExperimental Setup=Self-decided2026.05 | 26 | 458 |