Artificial Analysis

Artificial Analysis AutomationBench

An independently evaluated automation benchmark from Artificial Analysis.

In index
Agentic
Best in agentic
Source ↗

Models23

Top result77.8%Gemini 4 Carbon

Top-3 spread1.4 pts

Year2026

Result and price

21 of 23 results

Results

23 results

#ModelResult

01Gemini 4 Carbon77.8%

02Gemini 4 Argon77.5%

03Claude Fable 5.276.4%

04Kimi K3.1 Pro73.3%

05Claude Sonnet 5.571.8%

06Claude Opus 5.569.5%

07GPT-6 Astra68.5%

08Muse Ember68.1%

09Grok 4.765.6%

10GPT-6.1 Sol64.9%

11GLM-5.362.2%

12GPT-6.1 Terra60.2%

13Gemini 3.8 Flash59.9%

14Claude Fable 5.159.4%

15MiMo-V2.6-Pro58.6%

16Mistral Large 457.8%

17GPT-6 Luna53.2%

18Beam 1.0 501B-A23B48.5%

19Qwen3.8-27B48.2%

20Inkling 2 Small45.6%

21Muse Wonder Full45.4%

22Muse Wonder39.1%

23Claude Haiku 5.535.4%