# OSWorld-Verified leaderboard

Agentic benchmark

> Every figure below is reproduced as its upstream source published it: nothing is modelled, estimated, interpolated or converted. `Unavailable` means no source published the value — it is never a zero. Each value carries its evidence state and the date it was observed.

Page: https://modelscale.dev/leaderboards/benchmarks/agentic-os-world-verified

Every model the catalog carries a published OSWorld-Verified value for, ranked by that value.

## Benchmark record

| Field | Value |
| --- | --- |
| Category | Agentic |
| Measure | Execution-based interactive task success |
| Tasks | 369 real-world computer tasks (361 when eight Google Drive tasks are excluded) |
| Difficulty | Multi-step desktop and cross-application workflows |
| Published by | [OSWorld](https://os-world.github.io/) |

OSWorld-Verified is the July 2025 repaired release of OSWorld's real-computer evaluation. It measures whether a model-agent system can finish desktop and web tasks from configured starting states, with success checked by execution-based evaluators.

## OSWorld-Verified ranking

38 models with a published OSWorld-Verified value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.

| Rank | Model | Provider | Execution-based interactive task success | Evidence |
| --- | --- | --- | --- | --- |
| 1 | [Qwen3.8 Max](https://modelscale.dev/models/qwen3-8-max) | Alibaba | 86.1 | Observed 2026-11-21 · source benchlm:models |
| 2 | [Claude Fable 5](https://modelscale.dev/models/claude-fable-5) | Anthropic | 85 | Observed 2026-11-21 · source benchlm:models |
| 2 | [Claude Mythos 5](https://modelscale.dev/models/claude-mythos-5) | Anthropic | 85 | Observed 2026-11-21 · source benchlm:models |
| 4 | [Qwen3.8-27B](https://modelscale.dev/models/qwen3-8-27b) | Alibaba | 84.3 | Observed 2026-11-21 · source benchlm:models |
| 5 | [Claude Opus 4.8](https://modelscale.dev/models/claude-opus-4-8) | Anthropic | 83.4 | Observed 2026-11-21 · source benchlm:models |
| 6 | [Gemini 3.6 Flash](https://modelscale.dev/models/gemini-3-6-flash) | Google | 83 | Observed 2026-11-21 · source benchlm:models |
| 7 | [Holo3-35B-A3B](https://modelscale.dev/models/holo3-35b-a3b) | H Company | 82.56 | Observed 2026-11-21 · source benchlm:models |
| 8 | [MiMo-V2.6-Pro](https://modelscale.dev/models/mimo-v2-6-pro) | Xiaomi | 82 | Observed 2026-11-21 · source benchlm:models |
| 9 | [Claude Sonnet 5](https://modelscale.dev/models/claude-sonnet-5) | Anthropic | 81.2 | Observed 2026-11-21 · source benchlm:models |
| 10 | [MiMo-V2.6-Flash](https://modelscale.dev/models/mimo-v2-6-flash) | Xiaomi | 80.8 | Observed 2026-11-21 · source benchlm:models |
| 10 | [Muse Spark 1.1](https://modelscale.dev/models/muse-spark-1-1) | Meta | 80.8 | Observed 2026-11-21 · source benchlm:models |
| 12 | [Holo3-122B-A10B](https://modelscale.dev/models/holo3-122b-a10b) | H Company | 78.85 | Observed 2026-11-21 · source benchlm:models |
| 13 | [GPT-5.5](https://modelscale.dev/models/gpt-5-5) | OpenAI | 78.7 | Observed 2026-11-21 · source benchlm:models |
| 14 | [Gemini 3.5 Flash](https://modelscale.dev/models/gemini-3-5-flash) | Google | 78.4 | Observed 2026-11-21 · source benchlm:models |
| 15 | [Claude Opus 4.7](https://modelscale.dev/models/claude-opus-4-7) | Anthropic | 78 | Observed 2026-11-21 · source benchlm:models |
| 15 | [Claude Opus 4.7 (Adaptive)](https://modelscale.dev/models/claude-opus-4-7-max) | Anthropic | 78 | Last good 2026-11-18 · source benchlm:models |
| 17 | [UI-Mate-27B](https://modelscale.dev/models/ui-mate-27b) | Tencent | 77 | Observed 2026-11-21 · source benchlm:models |
| 18 | [GPT-5.4](https://modelscale.dev/models/gpt-5-4) | OpenAI | 75 | Observed 2026-11-21 · source benchlm:models |
| 19 | [Gemini 3.5 Flash-Lite](https://modelscale.dev/models/gemini-3-5-flash-lite) | Google | 74 | Observed 2026-11-21 · source benchlm:models |
| 20 | [Qwen3.7 Plus](https://modelscale.dev/models/qwen3-7-plus) | Alibaba | 73.3 | Observed 2026-11-21 · source benchlm:models |
| 21 | [Kimi K2.6](https://modelscale.dev/models/kimi-2-6) | Moonshot AI | 73.1 | Observed 2026-11-21 · source benchlm:models |
| 22 | [Claude Opus 4.6](https://modelscale.dev/models/claude-opus-4-6) | Anthropic | 72.7 | Observed 2026-11-21 · source benchlm:models |
| 23 | [Claude Sonnet 4.6](https://modelscale.dev/models/claude-sonnet-4-6) | Anthropic | 72.1 | Observed 2026-11-21 · source benchlm:models |
| 23 | [GPT-5.4 mini](https://modelscale.dev/models/gpt-5-4-mini) | OpenAI | 72.1 | Observed 2026-11-21 · source benchlm:models |
| 25 | [MiniMax M3](https://modelscale.dev/models/minimax-m3) | MiniMax | 70.06 | Observed 2026-11-21 · source benchlm:models |
| 26 | [Claude Opus 4.5](https://modelscale.dev/models/claude-opus-4-5) | Anthropic | 66.3 | Observed 2026-11-21 · source benchlm:models |
| 27 | [UI-Mate-9B](https://modelscale.dev/models/ui-mate-9b) | Tencent | 66.2 | Observed 2026-11-21 · source benchlm:models |
| 28 | [Muse Glimmer 30B](https://modelscale.dev/models/muse-glimmer-30b) | Meta | 65.9 | Observed 2026-11-21 · source benchlm:models |
| 29 | [GPT-5.3 Codex](https://modelscale.dev/models/gpt-5-3-codex) | OpenAI | 64.7 | Observed 2026-11-21 · source benchlm:models |
| 30 | [Kimi K2.5](https://modelscale.dev/models/kimi-k2-5) | Moonshot AI | 63.3 | Observed 2026-11-21 · source benchlm:models |
| 30 | [Kimi K2.5 (Reasoning)](https://modelscale.dev/models/kimi-k2-5-reasoning) | Moonshot AI | 63.3 | Last good 2026-11-18 · source benchlm:models |
| 32 | [Qwen3.5 397B](https://modelscale.dev/models/qwen3-5-397b) | Alibaba | 62.2 | Observed 2026-11-21 · source benchlm:models |
| 33 | [Claude Sonnet 4.5](https://modelscale.dev/models/claude-sonnet-4-5) | Anthropic | 61.4 | Observed 2026-11-21 · source benchlm:models |
| 34 | [Qwen3.5-122B-A10B](https://modelscale.dev/models/qwen3-5-122b-a10b) | Alibaba | 58 | Observed 2026-11-21 · source benchlm:models |
| 35 | [Qwen3.5-27B](https://modelscale.dev/models/qwen3-5-27b) | Alibaba | 56.2 | Observed 2026-11-21 · source benchlm:models |
| 36 | [Qwen3.5-35B-A3B](https://modelscale.dev/models/qwen3-5-35b-a3b) | Alibaba | 54.5 | Observed 2026-11-21 · source benchlm:models |
| 37 | [GPT-5.2](https://modelscale.dev/models/gpt-5-2) | OpenAI | 47.3 | Observed 2026-11-21 · source benchlm:models |
| 38 | [GPT-5.4 nano](https://modelscale.dev/models/gpt-5-4-nano) | OpenAI | 39 | Observed 2026-11-21 · source benchlm:models |

Rows are ordered by the value the source published, highest first. The source does not state whether a higher value is the better result, so this document does not either: for a benchmark that measures a rate of failure, read the table from the bottom.
