Toolathlon

A tool-use benchmark focused on selecting, sequencing, and completing tasks with external tools.

In index
Agentic

Models22

Top result75.6%Muse Spark 1.1

Top-3 spread17.6 pts

YearN/A

Result and price

Results

22 results

#ModelResult

01

Muse Spark 1.175.6%

02

Claude Opus 4.859.9%

03

GPT-5.6 Sol58.0%

04

Gemini 3.5 Flash56.5%

05

GPT-5.555.6%

06

GPT-5.454.6%

07

GPT-5.6 Luna53.4%

08

GPT-5.6 Terra53.1%

09

DeepSeek V4 Pro 081351.8%

10MAKimi K2.650.0%

11STStep 3.7 Flash49.5%

12

GLM-5.248.2%

13

DeepSeek V4 Flash 073147.8%

14

MiniMax M2.746.3%

15

Claude Opus 4.543.5%

16

GPT-5.4 mini42.9%

17

Qwen3.6 Plus39.8%

18

GLM-538.0%

19

Qwen3.5 397B36.3%

20

GPT-5.4 nano35.5%

21MAKimi K2.527.8%

22

Qwen3.6-35B-A3B26.9%