Function Calling on API-Bank Level-2
83.2ROUGE-LCoALM 70B
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| CoALM 70BParameters=70B2025.02 | 83.2 | 82.7 | 78.9 | 72.4 | — | |
| CoALM 8BParameters=8B2025.02 | 81.9 | 81.2 | 76.4 | 69.7 | — | |
| CoALM 405BParameters=405B, Training Status=One completed epoch2025.02 | 77.8 | 77.1 | 71.9 | 64.4 | — | |
| Llama 3.1 8B InstructParameters=8B, Mode=Instruct2025.02 | 75.2 | 81.4 | 76.3 | 65.1 | — | |
| Hammer 2.0 7BParameters=7B2025.02 | 74 | 74.1 | 68.5 | 58.4 | — | |
| Qwen2.5 7B InstructParameters=7B, Mode=Instruct2025.02 | 73.9 | 78.5 | 71.2 | 64.2 | — | |
| ToolAce2025.02 | 63.6 | 71.3 | 63 | 67 | — | |
| Granite-20B-CodeParameters=20B2025.02 | 45.7 | 48.9 | 43.4 | 29.3 | — | |
| LDST2025.02 | 7.1 | 11.6 | 2.4 | 5.7 | — | |
| tod-zero-bqag3oyb2025.02 | 4.2 | 12.4 | 2.2 | 0.9 | — | |
| Fnc-TOD 13BParameters=13B2025.02 | 3.3 | 23.4 | 9.2 | 1.1 | — | |
| nc-latent-tod-step-22025.02 | 3.2 | 13.3 | 1.5 | 0.8 | — | |
| Hammer-1.5BModel Size=1.5B2026.04 | — | — | — | — | 59.71 | |
| Hammer-7BModel Size=7B2026.04 | — | — | — | — | 65.98 | |
| Qwen2.5-1.5BModel Size=1.5B2026.04 | — | — | — | — | 42.93 | |
| Qwen2.5-32BModel Size=32B2026.04 | — | — | — | — | 63.04 | |
| Qwen2.5-7BModel Size=7B2026.04 | — | — | — | — | 60.36 | |
| two-step training model (AugFC)Model Size=1.5B2026.04 | — | — | — | — | 60.48 | |
| two-step training model (AugFC)Model Size=7B2026.04 | — | — | — | — | 64.4 | |
| two-step training model (AugFC)Model Size=32B2026.04 | — | — | — | — | 65.01 | |
| xLAM-1.3B-fcModel Size=1.5B2026.04 | — | — | — | — | 64.32 | |
| xLAM-2-32B-fcModel Size=32B2026.04 | — | — | — | — | 60 | |
| xLAM-7B-fcModel Size=7B2026.04 | — | — | — | — | 64.24 |