Function Calling on BFCL Non-Live
86.85Simple AccuracyGENESISFUNC-8B
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| GENESISFUNC-8Bbase_model=Qwen3-8B2026.04 | 86.85 | 97.25 | 96.15 | 93 | 93.31 | |
| Qwen-ToolRL-8B2026.04 | 83.5 | 89.5 | 91.25 | 91.85 | 89.03 | |
| ToolACE-8B2026.04 | 81.17 | 96 | 94 | 93 | 91.04 | |
| GPT-4o-mini2026.04 | 80.33 | 92 | 89.5 | 90.5 | 88.08 | |
| Qwen3-32B2026.04 | 79.58 | 95 | 92 | 93 | 89.9 | |
| Qwen3-8B2026.04 | 78.92 | 95 | 91.5 | 89 | 88.6 | |
| Gemini-2.5-Pro2026.04 | 78.67 | 94 | 93.5 | 92 | 89.54 | |
| LLaMA-3.1-70B-Instruct2026.04 | 78.33 | 97 | 95.5 | 94 | 91.21 | |
| Hammer2.1-7B2026.04 | 77.17 | 96 | 93 | 87.5 | 88.42 | |
| DeepSeek-V32026.04 | 76.58 | 94.5 | 92 | 92 | 87.77 | |
| GPT-4o2026.04 | 76.5 | 91 | 90 | 78 | 83.88 | |
| xLAM-8B2026.04 | 73.83 | 93.5 | 87.5 | 83.5 | 84.58 | |
| LLaMA-3.1-8B-Instruct2026.04 | 71 | 95 | 87.5 | 82.5 | 84 |