# OSWorld-Verified — AI model rankings

Can the AI actually operate a computer — clicking, typing, and using real apps — to finish tasks on its own? Higher is better.

16 tracked models have a published OSWorld-Verified score. Higher is better. Scores are as published at each model's release.

## Ranking

| Rank | Model | Developer | Score | Released |
| --- | --- | --- | --- | --- |
| 1 | Qwen3.8-Max | Qwen | 86.1% | Aug 3 2026 |
| 2 | Claude Fable 5 | Anthropic | 85% | Jun 9 2026 |
| 3 | Claude Opus 4.8 | Anthropic | 83.4% | May 28 2026 |
| 4 | Gemini 3.6 Flash | Google | 83% | Jul 21 2026 |
| 5 | Claude Sonnet 5 | Anthropic | 81.2% | Jun 30 2026 |
| 6 | Muse Spark 1.1 | Meta | 80.8% | Jul 9 2026 |
| 7 | GPT-5.5 | OpenAI | 78.7% | Apr 23 2026 |
| 8 | Gemini 3.5 Flash | Google | 78.4% | May 19 2026 |
| 9 | Claude Opus 4.7 | Anthropic | 78% | Apr 16 2026 |
| 10 | Gemini 3.1 Pro | Google | 76.2% | Feb 19 2026 |
| 11 | GPT-5.4 | OpenAI | 75% | Mar 5 2026 |
| 12 | Gemini 3.5 Flash-Lite | Google | 74% | Jul 21 2026 |
| 13 | Claude Sonnet 4.6 | Anthropic | 72.5% | Feb 17 2026 |
| 14 | Gemini 3.0 Flash | Google | 65.1% | Dec 17 2025 |
| 15 | Qwen3.5 | Qwen | 62.2% | Feb 16 2026 |
| 16 | Muse Spark | Meta | 53.3% | Apr 8 2026 |


---

Canonical page: https://aireleasetracker.com/benchmark/osworld-verified
Full dataset: https://aireleasetracker.com/llms-full.txt · JSON: https://aireleasetracker.com/models.json
Source: AI Release Tracker (https://aireleasetracker.com). Benchmark scores are the figures published by the releasing lab at launch.