# OSWorld 2.1 offline subset (82 tasks) (OSWorld 2.1 Offline)

> The official offline subset of OSWorld 2.1: 82 of the 108 long-horizon computer-use workflows, run with no internet access in the agent's virtual machine.

Canonical page: https://benchlm.ai/benchmarks/osworld-2-1-offline

- Category: [Agentic](/agentic)
- Last updated: October 29, 2026

## About OSWorld 2.1 Offline

- Year: 2026
- Tasks: 82 offline computer-use workflows
- Format: Interactive computer-use evaluation; partial score
- Difficulty: Long-horizon professional workflows
- Paper: [OSWorld 2.1 release (xlang-ai/OSWorld-V2, tag osworld-v2.1)](https://github.com/xlang-ai/OSWorld-V2/releases)

The osworld-v2.1 release added offline task subsets. Anthropic and OpenAI now report this 82-task offline list. We store the partial score (mean per-task checkpoint credit) and keep strict pass rates in the source notes. Scores are not comparable with the full 108-task OSWorld 2.1 page or with OSWorld 2.0.

OSWorld 2.1 Offline is currently displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Leaderboard (8 models)

| Rank | Model | Creator | Score |
|------|-------|---------|-------|
| 1 | [Claude Opus 5.5](/models/claude-opus-5-5) | Anthropic | 87.2% |
| 2 | [Claude Sonnet 5.5](/models/claude-sonnet-5-5) | Anthropic | 83.9% |
| 3 | [Gemini 4 Carbon](/models/gemini-4-carbon) | Google | 81.5% |
| 4 | [Gemini 4 Argon](/models/gemini-4-argon) | Google | 74.7% |
| 5 | [GPT-6.1 Terra](/models/gpt-6-1-terra) | OpenAI | 74.2% |
| 6 | [Claude Haiku 5.5](/models/claude-haiku-5-5) | Anthropic | 72.4% |
| 7 | [GPT-6 Luna](/models/gpt-6-luna) | OpenAI | 48.9% |
| 8 | [Claude Haiku 4.5](/models/claude-haiku-4-5) | Anthropic | 15.7% |

## FAQ

### What does OSWorld 2.1 Offline measure?

The official offline subset of OSWorld 2.1: 82 of the 108 long-horizon computer-use workflows, run with no internet access in the agent's virtual machine.

### Which model scores highest on OSWorld 2.1 Offline?

Claude Opus 5.5 by Anthropic currently leads with a score of 87.2% on OSWorld 2.1 Offline.

### How many models are evaluated on OSWorld 2.1 Offline?

8 AI models have published results on OSWorld 2.1 Offline in the BenchLM catalog.

### Does OSWorld 2.1 Offline affect BenchLM's overall score?

Not directly. OSWorld 2.1 Offline is still displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Compare Top Models on OSWorld 2.1 Offline

- [Claude Opus 5.5 vs Claude Sonnet 5.5](/compare/claude-opus-5-5-vs-claude-sonnet-5-5)
- [Claude Sonnet 5.5 vs Gemini 4 Carbon](/compare/claude-sonnet-5-5-vs-gemini-4-carbon)
- [Gemini 4 Carbon vs Gemini 4 Argon](/compare/gemini-4-argon-vs-gemini-4-carbon)
- [Gemini 4 Argon vs GPT-6.1 Terra](/compare/gemini-4-argon-vs-gpt-6-1-terra)