benchmark evidence
BFCL Overall
BFCL overall function calling accuracy.
winner on BFCL Overall
direct benchmark result, not a broad vertical composite | source row dated 2026-05-15
scored on 2026-05-15
latest mapped results | top 20
| # | Model | Score | Evidence | Tested |
|---|---|---|---|---|
| 1 | MoonshotAI: Kimi K2 Thinking | 59.1 | model-only independent_benchmark | 2026-05-15 |
| 2 | DeepSeek: DeepSeek V3.2 | 54.1 | model-only independent_benchmark | 2026-05-15 |
| 3 | Google: Gemini 2.5 Flash | 50.9 | model-only independent_benchmark | 2026-05-15 |
| 4 | OpenAI: o4 Mini | 50.3 | model-only independent_benchmark | 2026-05-15 |
| 5 | OpenAI: o3 | 48.6 | model-only independent_benchmark | 2026-05-15 |
| 6 | OpenAI: GPT-5.2 | 45.3 | model-only independent_benchmark | 2026-05-15 |
| 7 | OpenAI: GPT-4.1 | 39.4 | model-only independent_benchmark | 2026-05-15 |
| 8 | Meta: Llama 4 Maverick | 37.3 | model-only independent_benchmark | 2026-05-15 |
| 9 | Anthropic: Claude Opus 4.5 | 33.5 | model-only independent_benchmark | 2026-05-15 |
| 10 | Mistral: Mistral Small 4 | 32.4 | model-only independent_benchmark | 2026-05-15 |
| 11 | Meta: Llama 3.3 70B Instruct | 31.9 | model-only independent_benchmark | 2026-05-15 |
| 12 | Mistral Large | 31.8 | model-only independent_benchmark | 2026-05-15 |
| 13 | Google: Gemma 3 27B | 29.5 | model-only independent_benchmark | 2026-05-15 |
| 14 | Meta: Llama 4 Scout | 28.1 | model-only independent_benchmark | 2026-05-15 |
| 15 | Anthropic: Claude Haiku 4.5 | 25.3 | model-only independent_benchmark | 2026-05-15 |
| 16 | Amazon: Nova Pro 1.0 | 25.0 | model-only independent_benchmark | 2026-05-15 |
| 17 | Anthropic: Claude Sonnet 4.5 | 24.9 | model-only independent_benchmark | 2026-05-15 |
what this result means
BFCL overall function calling accuracy.
This benchmark contributes direct public evidence. Read its scope before generalizing the result.
A win here is a win on BFCL Overall. Broad task pages require independent corroboration before naming a general winner.
source record
category: structured_output
metric: accuracy
matched models: 17
latest source date: 2026-05-15
direction: higher is better