Atlas

Benchmarks

← All benchmarks

AutomationBench 1.0.6 (provider-reported)

Agents

Model-provider evaluation identifying AutomationBench version 1.0.6 but not the complete task roster or public/private split. Strict workflow success is a component of the Zapier completion family; it is not Artificial Analysis's partial-objective metric.

Top models (higher is better)

ModelScore
DeepSeek V4.1 Flash 101760.5
DeepSeek V4.1 Flash54.8
Atria Dawn Preview53.8
MiMo-V2.6-Pro53.1
MiMo-V2.6-Flash52.3
GLM-5.348.2
Loading Atlas data…