BENCHMARKS UC Berkeley Expands BFCL Benchmark to v4, Adding Agentic Evaluation for LLM Function Calling 4/10 2 min read 2 months ago