LLM Reference
activeAgents

AutomationBench

Metric: Accuracy (higher is better)Introduced: 2026

Multi-step computer automation and workflow-completion benchmark for agentic UI tasks.

Models ranked

6

tracked on this benchmark

Score band

54.8 – 12.9

best → lowest tracked

Snapshot trend

+27.50

Aug 12 → Sep 10 · 1 models

Leaderboard

Tracked models ranked by Accuracy (higher is better).

Compare candidates
#Model variant and provenanceScore
1
DeepSeek V4.1 Flash
Version: AutomationBench Pass@1; official scaffold; max reasoning effortHarness: Not recordedEvaluator: Not recordedObserved: Sep 10, 2026Confidence: Not recordedSource
54.8
2
Kimi K3
Version: 600-task public subset; official GitHub setup; max reasoning effortHarness: Not recordedEvaluator: Not recordedObserved: Jul 16, 2026Confidence: Not recordedSource

Notes: Model variant: Kimi K3 (max reasoning effort). Benchmark variant: AutomationBench 600-task public subset. Harness/evaluator: official AutomationBench GitHub setup; evaluator not disclosed by Moonshot. Source posture: Moonshot/Kimi self-reported launch table; confidence: medium because the value is vendor-reported and not independently reproduced. Recommended seed value: 30.8.

30.8
3
Gemini 3.7 Flash
Version: AutomationBench private setHarness: Not recordedEvaluator: Not recordedObserved: Aug 13, 2026Confidence: Not recordedSource
30.4
4
Qwen3.8-Max
Version: Automation-Bench (Pass@1)Harness: Not recordedEvaluator: Not recordedObserved: Aug 12, 2026Confidence: Not recordedSource
27.3
5
Claude Fable 5
Version: AutomationBenchHarness: Not recordedEvaluator: Not recordedObserved: Jun 9, 2026Confidence: Not recordedSource

Notes: DAT-6106: Unstarred in Anthropic's shared Fable 5/Mythos 5 launch table, so this is a Claude Fable 5 score.

17.4
6
GPT-5.5
Version: AutomationBenchHarness: Not recordedEvaluator: Not recordedObserved: Jun 9, 2026Confidence: Not recordedSource

Notes: DAT-6106: GPT-5.5 score from Anthropic's Fable 5 comparison table.

12.9

How to read this benchmark

This benchmark scores models where higher is better. Use scores for directional filtering and shortlisting, not universal quality ranking; then validate pricing, context window, provider availability, and fit for your workload.

Trust this score when

  • There is a fresh timestamped snapshot (or multiple snapshots) for this benchmark.
  • The model list covers the same version family you can actually deploy today.
  • Top candidates overlap with your required routing and feature requirements.

Be cautious when

  • There is only one benchmark snapshot or the dataset appears stale.
  • The benchmark metric direction is opposite of your decision objective.
  • The score difference between options is narrow and likely within implementation variance.

Related benchmarks

Last reviewed: Jun 15, 2026

Resources