Function Calling on API-Bank Level-1
93.4ROUGE-LCoALM 405B
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| CoALM 405BParameters=405B, Training Status=One completed epoch2025.02 | 93.4 | 94.5 | 92.4 | 90.3 | — | |
| CoALM 8BParameters=8B2025.02 | 92.8 | 94.1 | 91.9 | 89.4 | — | |
| CoALM 70BParameters=70B2025.02 | 92.7 | 94.5 | 92.5 | 89.5 | — | |
| Hammer 2.0 7BParameters=7B2025.02 | 90.1 | 92.3 | 89.9 | 85.4 | — | |
| Qwen2.5 7B InstructParameters=7B, Mode=Instruct2025.02 | 84.3 | 88.9 | 84.6 | 76.4 | — | |
| ToolAce2025.02 | 81.5 | 88.8 | 85 | 76.1 | — | |
| Llama 3.1 8B InstructParameters=8B, Mode=Instruct2025.02 | 72.7 | 84 | 79.8 | 62.3 | — | |
| Granite-20B-CodeParameters=20B2025.02 | 60.3 | 64.7 | 59.5 | 43.8 | — | |
| LDST2025.02 | 8.3 | 12.8 | 2.7 | 6.2 | — | |
| Fnc-TOD 13BParameters=13B2025.02 | 3.9 | 22.1 | 8 | 1.5 | — | |
| tod-zero-bqag3oyb2025.02 | 3.7 | 11.5 | 1.1 | 1 | — | |
| nc-latent-tod-step-22025.02 | 3.2 | 14.3 | 3.2 | 0.8 | — | |
| Hammer-1.5BModel Size=1.5B2026.04 | — | — | — | — | 72.3 | |
| Hammer-7BModel Size=7B2026.04 | — | — | — | — | 83.11 | |
| Qwen2.5-1.5BModel Size=1.5B2026.04 | — | — | — | — | 65.25 | |
| Qwen2.5-32BModel Size=32B2026.04 | — | — | — | — | 79.88 | |
| Qwen2.5-7BModel Size=7B2026.04 | — | — | — | — | 69.85 | |
| two-step training model (AugFC)Model Size=1.5B2026.04 | — | — | — | — | 72.26 | |
| two-step training model (AugFC)Model Size=7B2026.04 | — | — | — | — | 80.02 | |
| two-step training model (AugFC)Model Size=32B2026.04 | — | — | — | — | 84.16 | |
| xLAM-1.3B-fcModel Size=1.5B2026.04 | — | — | — | — | 83.7 | |
| xLAM-2-32B-fcModel Size=32B2026.04 | — | — | — | — | 82.7 | |
| xLAM-7B-fcModel Size=7B2026.04 | — | — | — | — | 80.69 |