multimodal benchmark
OSWorld Screenshot-only Leaderboard
OSWorld Screenshot-only is a multimodal benchmark for evaluating AI agents that complete open-ended computer tasks using only screenshot observations across Ubuntu, Windows, and macOS.
Updated Sep 5, 2026
Models1
Model coverage1
MetricScore
EvidenceB
OSWorld Screenshot-only Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelAN | Score14.90% | Percentile100.00% | Participants1 | EvidenceC | Evaluated |
OSWorld Screenshot-only Highlights
The leading models and scores on this benchmark.
The Top AI Models for OSWorld Screenshot-only
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
Ranking basisThis osworld screenshot-only AI model leaderboard uses descending score in the benchmark original unit. The leaderboard ranking keeps matched price and speed data separate from benchmark evidence.
- 01ANAnthropic
- Score
- 14.90%
- Speed
- Up to 100.00 tok/s via Anthropic
Strengths
- Ranks #1 of 1 compared models
- 100th percentile on this benchmark
- C evidence result
Considerations
- This result measures OSWorld Screenshot-only, not total model capability
Selection summary
Best AI Models for OSWorld Screenshot-only
Claude 3.5 Sonnet currently leads OSWorld Screenshot-only with 14.90%. It is the top model on this specific benchmark, while the best LLM for the broader task should also be checked against other benchmarks, price, and runtime.
Use this leaderboard with the supporting benchmark results and coverage details above. A leaderboard position summarizes the selected ranking signal; it does not replace workload-specific testing.
What is OSWorld Screenshot-only?
What OSWorld Screenshot-only measures and how its scores work.
A variant of the OSWorld benchmark in which multimodal AI agents complete open-ended tasks across real operating systems using only screenshot observations.
It measures agents' ability to complete workflows involving web apps, desktop applications, file I/O, and multi-application tasks through visual interface understanding and GUI grounding, reported as a Score ratio.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- OSWorld Screenshot-only
- Modality
- multimodal
- Primary category
- multimodal
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- overall
LLM Stats. Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about OSWorld Screenshot-only.
Which model scores highest on OSWorld Screenshot-only?
Claude 3.5 Sonnet is currently ranked first with 14.90%.
What are the top three models on OSWorld Screenshot-only?
The current leaders are Claude 3.5 Sonnet (14.90%).
Which OSWorld Screenshot-only model has the lowest official input price?
No matched official input price is currently available.
Which models are fastest among OSWorld Screenshot-only results?
The fastest matched records are Claude 3.5 Sonnet (100.00 tok/s via Anthropic).
Does the highest score result prove overall model quality?
No. This benchmark measures one defined capability or task. The overall LLMBoard score uses a separate aggregation across eligible benchmark evidence.