UnifyBench

RankingAutomationBench-AA

AutomationBench-AA

Data updated 31 Oct 2026

Bucket
Supporting evidence
Unit
percent
Direction
Higher is better
Version
aa-score
Display harness
Artificial Analysis AutomationBench-AA
Board
https://artificialanalysis.ai/evaluations/automationbench-aa

AutomationBench-AA score: share of task objectives completed with no guardrail violation, on Artificial Analysis's private 657-task holdout. This is not Zapier's tasks-completed rate. The effort dataset already includes this evaluation in the agentic family under the existing equal-family budget. These rows are the board, not a second weight.

Compare published benchmark results with category weights →

Models

Results are ordered by score in the display harness, followed by models with no result. Other harnesses are listed separately below. Missing results remain unknown.

51–75 of 471 entries

ModelScoreEvidenceSource-recorded date
Ministral 3 3BMistral0.6658657772309807%official board2026-10-30
Ministral 3 14BMistral0.5512132673736124%official board2026-10-30
Ministral 3 8BMistral0.516749466666553%official board2026-10-30
Solar Pro 3Upstage0.49734242022629577%official board2026-10-30
Mistral Small 3.2Mistral0.2854023922677291%official board2026-10-30
Llama 4 MaverickMeta0.2121918776569101%official board2026-10-30
gpt-oss-120bOpenAI0.19906990691605594%official board2026-10-30
gpt-oss-20bOpenAI0.19111706206154658%official board2026-10-30
Llama-4-Scout-17B-16E-InstructMeta0.19111706206154658%official board2026-10-30
Magistral Small 1.2Mistral0.19111706206154658%official board2026-10-30
Agnes-3.0-Flash PreviewAgnes AI———
Amazon Nova 2 LiteAmazon———
Amazon Nova 2 Pro PreviewAmazon———
Amazon Nova LiteAmazon———
Amazon Nova MicroAmazon———
Amazon Nova PremierAmazon———
Amazon Nova ProAmazon———
AutoGLM-Phone-9BZ.ai———
AutoGLM-Phone-9B-MultilingualZ.ai———
aya-101Cohere———
aya-23-35BCohere———
aya-23-8BCohere———
aya-expanse-8bCohere———
aya-vision-8bCohere———
C4Ai Aya Expanse 32BCohere———