Galileo Tool Tasks - tau Long Context: leaderboard

Metric: Accuracy (%). Source: huggingface.co. 19 models tracked.

Top models

#ModelScore
1Claude 3.7 Sonnet (20250219)100
2O1 (2024-12-17)98
3Qwen 2.5 72B Instruct92
4Mistral Small 392
5GPT-4o (2024-11-20)92
6Gemini 2.0 Flash (001)90
7Gemini 1.5 Flash89
8O3 Mini (2025-01-31)84
9Mistral Large 2 (Nov) Instruct (2411)83
10Gemini 1.5 Pro75
11Ministral-8B-Instruct-241073
12Llama 3.3 70B Instruct72
13Claude 3.5 Haiku (20241022)72
14Claude 3.5 Sonnet (20241022)68
15GPT-4o Mini51

Interactive version: theaggregate.ai/benchmark?slug=galileo-tool-tasks-tau-long-context · How It Works · Data refreshed daily, snapshot 2026-09-05.