Skip to content

Terminal-Bench-Science 0.1 with sixfold verifier timeout

Google-run scientific workflow evaluation using a sixfold verifier timeout to address verification timeouts.

1 models

Higher is better.

  1. 1Gemini 4 ArgonGoogle57.6
Higher is better. Bar colour: blue proprietary, orange open weights, green other.

About this benchmark

Metric
Resolution rate across 3 trials per task
Tasks
70 expert-curated scientific research workflows
Authors
Google DeepMind (2026)

Scores are as reported by model developers or evaluators and may use different settings (for example reasoning effort or tools). Compare with care.