Wikibench

BFCL-V4

agents

Multilingual: no

Berkeley Function Calling Leaderboard V4 (BFCL-V4) evaluates LLMs on their ability to accurately call functions and APIs, including simple, multiple, parallel, and nested function calls across diverse programming scenarios.

Methodology

Imported from llm-stats public benchmark metadata. Modality: text. Max score: 1. Categories: agents, tool_calling. Language: en. Verified by llm-stats: no.

Leaderboard

Showing 8 of 8 results

  1. Qwen3.7 Max self-reported llm-stats
    75.0%
  2. Qwen3.5-397B-A17B self-reported llm-stats
    72.9%
  3. Qwen3.5-122B-A10B self-reported llm-stats
    72.2%
  4. Qwen3.5-27B self-reported llm-stats
    68.5%
  5. Qwen3.5-35B-A3B self-reported llm-stats
    67.3%
  6. Nova 2 Pro self-reported llm-stats
    61.6%
  7. Nova 2 Lite self-reported llm-stats
    60.3%
  8. Nova 2 Omni self-reported llm-stats
    58.3%