BFCL V4
Berkeley Function Calling Leaderboard version 4, a benchmark for evaluating the tool-calling capabilities of large language models. It includes multi-turn interaction task scenarios.
Timeline (last 90 days)
In 120 function-calling tests on Underdog Bench 120—built on Berkeley's BFCL V4 leaderboard—Saluki 27B scored 88, surpassing the full-size original model's score of 84.
The researchers selected 159 multi-turn BFCL V4 tasks with complete precise replay paired data for the main analysis.
All Facts (2)
In 120 function-calling tests on Underdog Bench 120—built on Berkeley's BFCL V4 leaderboard—Saluki 27B scored 88, surpassing the full-size original model's score of 84.
50%UnverifiedThe researchers selected 159 multi-turn BFCL V4 tasks with complete precise replay paired data for the main analysis.
50%