BFCL V4 - Irrelevance Detection: leaderboard
Berkeley Function Calling Leaderboard version 4 evaluates API and tool use. The selected task column measures overall, non-live AST, live, multi-turn, web-search, memory, relevance or irrelevance performance as an accuracy percentage. This board reports BFCL V4 - Irrelevance Detection using Accuracy (%).
Metric: Accuracy (%). Source: gorilla.cs.berkeley.edu. Saturation forecast: Around February 2027. 109 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Mistral Medium 3 | 74.49 | #507 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=bfcl-v4-irrelevance-detection · How It Works · Data refreshed daily, snapshot 2026-11-05.