Function Calling on Berkeley Function-Calling Leaderboard (BFCL) (AST Success Rates)
96Non-Live Multiple AST Success RateBielik-11B-v3.0-Instruct (FC)
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Bielik-11B-v3.0-Instruct (FC)Params (B)=11.02026.03 | 96 | 88 | 82 | 79.07 | 72.36 | 62.5 | 75 | |
| Gemma-3-12b-it (Prompt)Params (B)=12.02026.03 | 95 | 90 | 73 | 84.88 | 70.85 | 87.5 | 62.5 | |
| Bielik-Minitron-7BParams (B)=7.32026.03 | 94.5 | 92 | 85 | 71.32 | 71.51 | 31.25 | 66.67 | |
| Bielik-11B-v2.6-InstructParams (B)=11.02026.03 | 94.5 | 87.5 | 86 | 75.97 | 76.07 | 37.5 | 66.67 | |
| Bielik-11B-v2.3-Instruct (Prompt)Params (B)=11.02026.03 | 93.5 | 47 | 50 | 72.87 | 69.71 | 43.75 | 54.17 | |
| Open-Mistral-Nemo-2407Params (B)=12.22026.03 | 93.5 | 85.5 | 85 | 77.13 | 69.61 | 75 | 70.83 | |
| Bielik-4.5B-v3.0-Instruct (FC)Params (B)=4.62026.03 | 92.5 | 82 | 86 | 70.16 | 68.66 | 50 | 54.17 |