agents benchmark
BFCL-V4
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.
Updated Aug 7, 2026
Published models13
Registry coverage13
MetricScore
EvidenceB
BFCL-V4 leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | AC | 75.0% | 100.0% | 13 | C | |
| 2 | AC | 72.9% | 91.7% | 13 | C | |
| 3 | AC | 72.9% | 83.3% | 13 | C | |
| 4 | AC | 72.2% | 75.0% | 13 | C | |
| 5 | AC | 68.5% | 66.7% | 13 | C | |
| 6 | AC | 67.3% | 58.3% | 13 | C | |
| 7 | AC | 66.1% | 50.0% | 13 | C | |
| 8 | AM Nova 2 ProAmazon | 61.6% | 41.7% | 13 | C | |
| 9 | AMNova 2 LiteAmazon | 60.3% | 33.3% | 13 | C | |
| 10 | AMNova 2 OmniAmazon | 58.3% | 25.0% | 13 | C | |
| 11 | AC | 50.3% | 16.7% | 13 | C | |
| 12 | AC | 43.6% | 8.3% | 13 | C | |
| 13 | AC | 25.3% | 0.0% | 13 | C |
Score distribution
Top published rows on the benchmark's original scale.
BFCL-V4 highlights
The top published results on this benchmark's own scale.
What is BFCL-V4?
Definition and scoring fields from the benchmark registry.
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.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- BFCL-V4
- Modality
- text
- Primary category
- agents
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- bfcl-v4|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about BFCL-V4.
Which model scores highest on BFCL-V4?
Qwen3.7 Max is currently ranked first with 75.0%.
What does BFCL-V4 measure?
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.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
13 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is preserved as source-native evidence but is not eligible for the current overall score.