BlogLLM Benchmark Scores as Published, Side by Side
BFCL: Results as Published
Every result the catalog holds for BFCL: 16 models from 8 developers, read from 5 pages. Each number is as printed by the company named beside it; nothing is ranked or rescored. Models are listed by developer, newest first.
Area Tool Use and WorkflowsMaintainer UC Berkeley Gorilla project (Shishir G. Patil and co-authors)Introduced 2024Data version 2026103001, as of 2026-10-30
The Berkeley Function Calling Leaderboard (BFCL) tests whether a model produces correct function calls from a tool description and a request. Version 3 added multi-turn interactions checked by the resulting system state; version 4 added agentic evaluation. Overall accuracy is the unweighted average of the sub-categories.
BFCL v4
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-5.2OpenAI, released 2025-12-11 | 63.1% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| GPT-5 miniOpenAI, released 2025-08-07 | 55.5% | Alibaba, Qwen3.5-122B-A10B model cardchecked 2026-10-15 | - |
| Claude Opus 4.5Anthropic, released 2025-11-24 | 77.5% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| Gemini 3 ProGoogle, released 2025-11-18 | 72.5% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| Qwen3.5-122B-A10BAlibaba, released 2026-02-24 | 72.2% | Alibaba, Qwen3.5-122B-A10B model cardchecked 2026-10-15 | - |
| Qwen3.5-27BAlibaba, released 2026-02-24 | 68.5% | Alibaba, Qwen3.5-122B-A10B model cardchecked 2026-10-15 | - |
| Qwen3.5-35B-A3BAlibaba, released 2026-02-24 | 67.3% | Alibaba, Qwen3.5-122B-A10B model cardchecked 2026-10-15 | - |
| Qwen3.5-397B-A17BAlibaba, released 2026-02-16 | 72.9% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| Qwen3-30B-A3B-Thinking-2507Alibaba, released 2025-07-30 | 46.4% | NVIDIA, NVIDIA Nemotron 3 Nano 30B-A3B model cardchecked 2026-10-15 | Effort: thinking |
| Qwen3-235B-A22BAlibaba, released 2025-04-29 | 54.8% | Alibaba, Qwen3.5-122B-A10B model cardchecked 2026-10-15 | - |
| Qwen3-Max-ThinkingAlibaba | 67.7% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| Kimi K2.5Moonshot AI, released 2026-01-27 | 68.3% | Alibaba, Qwen3.5-397B-A17B model cardchecked 2026-10-15 | - |
| Nemotron 3 Nano 30B-A3BNVIDIA, released 2025-12-15 | 53.8% | NVIDIA, NVIDIA Nemotron 3 Nano 30B-A3B model cardchecked 2026-10-15 | - |
BFCL v3 (accuracy)
accuracy
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| Claude Sonnet 4.6Anthropic, released 2026-02-17 | 76% | Microsoft, Microsoft AI: MAI-Thinking-1 technical reportchecked 2026-10-19 | Effort: max |
| MAI-Thinking-1Microsoft, released 2026-08-12 | 72% | Microsoft, Microsoft AI: MAI-Thinking-1 technical reportchecked 2026-10-19 | - |
BFCL v3 multi-turn
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| DeepSeek-R1-0528DeepSeek, released 2025-05-28 | 37% | DeepSeek, DeepSeek-R1-0528 model cardchecked 2026-10-15 | - |
Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. Where one page printed a model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.