Galileo Tool Tasks - BFCL v3 Multi-Turn Missing Function: leaderboard

Tool-use evaluation from Galileo covering BFCL v3 cases where the needed function is absent and models must respond appropriately.

Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 19 models tracked.

Top models

#ModelScoreOverall rank
1O1 (2024-12-17)94#149
2Gemini 2.0 Flash (001)93#360
3Gemini 1.5 Flash92#618
4Claude 3.7 Sonnet (20250219)92#201
5O3 Mini (2025-01-31)91#240
6Gemini 1.5 Pro86#445
7GPT-4o (2024-11-20)86#381
8Qwen 2.5 72B Instruct78#454
9Mistral Small 374#627
10GPT-4o Mini73#599
11Mistral Large 2 (Nov) Instruct (2411)70#465
12Claude 3.5 Sonnet (20241022)69#291
13Claude 3.5 Haiku (20241022)66#588
14Llama 3.3 70B Instruct54#534
15Ministral-8B-Instruct-241053#1094

No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.

Interactive version: theaggregate.ai/benchmark?slug=galileo-tool-tasks-bfcl-v3-multi-turn-missing-function · How It Works · Data refreshed daily, snapshot 2026-10-20.