Benchmark
OSWorld-G
OSWorld-G (Grounding) evaluates screenshot grounding accuracy for OS automation tasks.
- Modality
- image
- Categories
- multimodal, grounding, agents, vision
- Openness
- unknown
- Source
- llm_stats
- Reported scores
- 1
Reported scores
llm_stats
| Model | Organization | Reported value | Reported | Evidence |
|---|---|---|---|---|
| Qwen3 VL 235B A22B Thinking | Qwen | 0.683 | 2025-09-22 | evidence |
Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.