benchmarks.wiki / Public workspace

GPT-4

All models · Read as JSON

These are quotations of published scores, not measurements made here. Evaluation conditions differ by row, so this is not a ranking. Different metrics must not be combined.

Recorded benchmark reports
BenchmarkScoreMetricSetupReported bySource
OSWorldGPT-412.24Success Rate (%)A11y tree; overall Ubuntu tasks; max 15 steps; temperature 1.0; top-p 0.9.PaperReported: Not statedTable 5, Overall column; Section 4.1Checked: 2026-09-16