Versions

1 version · Metric: LABBench2 Accuracy

Versions of LABBench2
VersionReleasedModelsResults
LABBench2 Terminal and internet — September 2026 evaluationConfiguration: Terminal and internetSeptember 30, 202611

Recently reported results: LABBench2 Terminal and internet — September 2026 evaluation

Most recent configuration: LABBench2 Terminal and internet — September 2026 evaluation

Recently reported results for LABBench2 Terminal and internet — September 2026 evaluation
ModelScoreSource
Gemini 4 Argon (highest thinking settings)88.8%deepmind.google for Gemini 4 Argon (highest thinking settings) on LABBench2 Terminal and internet — September 2026 evaluation (opens in a new tab)LABBench2 AccuracyReported Evaluation methodologyGoogle DeepMindSelf-reportedRecord for Gemini 4 Argon on LABBench2 Terminal and internet — September 2026 evaluation