Benchmarks / Microsoft STATE-Bench

Reported by Microsoft STATE-Bench

Microsoft STATE-Bench

Share of tasks completed correctly, averaged over repeated runs.

Results dated
1 Jul 2026
Models
2
Unit
% of tasks
Licence
MIT License
STATE-Bench v0.8: task success (pass@1), % of tasks, higher is better
#ModelSTATE-Bench v0.8: task success (pass@1)
% of tasks, higher is better
1 GPT-5.5OpenAI · gpt-5.5
58.9%
2 GPT-5.4OpenAI · gpt-5.4
58.6%

Results as published by Microsoft STATE-Bench; we do not re-run them.

What it measures

Share of tasks completed correctly, averaged over repeated runs.

What it does not measure

Not the model alone: it runs inside STATE-Bench's agent loop against simulated users, and versions of the benchmark are not comparable.

Source

Microsoft STATE-Bench · Methodology

Failures

Failures count against a model: a product with no usable output is a failed listing. They are listed here so you can see why.

Not ranked

  • GPT 5.1: memory-track entries are agent-learning systems, not models
  • GPT 5.1 + Foundry Memory: memory-track entries are agent-learning systems, not models
  • GPT 5.4: memory-track entries are agent-learning systems, not models
  • GPT 5.4 + Foundry Memory: memory-track entries are agent-learning systems, not models

STATE-Bench by Microsoft and STATE-Bench contributors, MIT License.