RankingOSWorld partial · 2.0 pre-08.08

OSWorld partial · 2.0 pre-08.08

Data updated 12 Sept 2026

Bucket
Agentic
Unit
percent
Direction
Higher is better
Version
2.0 pre-08.08
Display harness
Gemini3.8Flash Model Card
Board
https://storage.googleapis.com/deepmind-media/Model-Cards/Gemini-3-8-Flash-Model-Card.pdf

The available records have no admitted matched comparison in the capability core. Raw results remain available below.

Compare published benchmark results with category weights →

Models

ModelScoreHarnessEvidenceSource-recorded date
Gemini 3.8 FlashGoogle59%
Reported settings & source

Partialscore; batchtools;1080p/500steps; Gemini/Sonnet bestof3runs; screenshotonly; officialCUAharness

Methodology says runs pre08.08patch but Opusvalue fromFable5.1card usesAugustfixedtasks; no controlledsameversionclaim. GPT values providerreports.

Gemini3.8Flash Model Card · Model card page5 Results table · reviewed 2026-09-06

Gemini3.8Flash Model Cardlab self-report2026-09-06
Gemini 3.7 FlashGoogle50.6%
Reported settings & source

Partialscore; batchtools;1080p/500steps; Gemini/Sonnet bestof3runs; screenshotonly; officialCUAharness

Methodology says runs pre08.08patch but Opusvalue fromFable5.1card usesAugustfixedtasks; no controlledsameversionclaim. GPT values providerreports.

Gemini3.8Flash Model Card · Model card page5 Results table · reviewed 2026-09-06

Gemini3.8Flash Model Cardlab self-report2026-09-06
Claude Sonnet 5Anthropic42.6%
Reported settings & source

Partialscore; batchtools;1080p/500steps; Gemini/Sonnet bestof3runs; screenshotonly; officialCUAharness

Methodology says runs pre08.08patch but Opusvalue fromFable5.1card usesAugustfixedtasks; no controlledsameversionclaim. GPT values providerreports.

Gemini3.8Flash Model Card · Model card page5 Results table · reviewed 2026-09-06

Gemini3.8Flash Model Cardlab self-report2026-09-06