RecentsView All

Your evaluations

Sign in to run benchmarks and follow your results here.

Benchgen

Overview Leaderboard Evaluations Phases

BFCL-v4 — Results

Rank	Model	Score
1	lfm2-5-2-6b	56.88
2	lfm2-5-vl-3b	32.5

B

BFCL-v4 — Leaderboard

1 phaseActive

BFCL-v4 Leaderboard

No self-reported entries.

About

Berkeley Function Calling Leaderboard v4 — adds holistic agentic evaluation (including web search) to AST-based function-calling accuracy.

Pages

Phases: 1Active
Participants: 1
Submissions: 2
Last Modified: Aug 4, 2026