OpenCodePapers

computer-use-agents-on-osworld-2-0

AgentsComputer Use Agents
Dataset Link
Results over time
Click legend items to toggle metrics. Hover points for model names.
Leaderboard
PaperCodeSuccess RateModelNameReleaseDate
Introducing Claude Opus 5✓ Link70.57Claude Opus 5 (Anthropic system card)2026-07-24
GLM-5.3-Flash: Frontier Intelligence, Flash Cost✓ Link59.1GLM-5.3-Flash2026-08-26
Kimi K3: Open Frontier Intelligence✓ Link58.3Kimi K3 Max2026-07-27
Introducing Muse Spark 1.332.0Muse Spark 1.3 (max)2026-09-02
StateAct: Program State, before Pixels, for Long-Horizon Computer-Use Agents26.9StateAct (Claude Opus 4.8)2026-07-24
Qwen-CUA: Native Computer Use for (almost) Everything✓ Link21.2Qwen-CUA-Max2026-08-02
Qwen3.8-Max: A New Bar for Coding and Cowork✓ Link19.4Qwen3.8-Max2026-08-03
Introducing Muse Spark 1.1✓ Link14.2Muse Spark 1.12026-07-09
LongHorizon-Harness: Advancing Long-Horizon Agents for Real-World Tasks✓ Link8.3Qwen 3.7-Plus + LongHorizon-Harness2026-08-03