OSWorld-Verified
OSWorld-Verified - computer use benchmark evaluating autonomous GUI operation on verified tasks across desktop OS environments.
Category
General Agents
Max Score
100.0
Score Type
percent
Active
Yes
Model Scores
Scores for OSWorld-Verified
| Model | Score | Date | Verified |
|---|---|---|---|
| GPT 5.5 |
87.60%
|
11.06.2026 | Unverified |
| Kimi K2.6 |
76.22%
|
11.06.2026 | Unverified |
| Claude Sonnet 4.6 |
75.00%
|
11.06.2026 | Unverified |
| Qwen3.8-27B |
98.98%
|
18.08.2026 | Verified |
| Qwen3.6-27B |
57.52%
|
18.08.2026 | Unverified |
| Qwen3.7-Plus |
76.63%
|
18.08.2026 | Unverified |
| Muse Glimmer-30B |
61.59%
|
18.08.2026 | Unverified |
| Opus4.6 Max |
75.41%
|
18.08.2026 | Unverified |
| Qwen3.5-4B |
35.60
|
01.02.2026 | Verified |
| Muse-Glimmer-30B |
61.59%
|
01.08.2026 | Verified |
| Gemma4-31B-it |
46.54%
|
01.08.2026 | Unverified |
| Claude Opus 4.7 |
95.94%
|
11.06.2026 | Unverified |
| MiniMax-M3 |
80.49%
|
11.06.2026 | Unverified |
| Gemini 3.1 Pro |
82.52%
|
11.06.2026 | Unverified |
| Qwen3.5-122B-A10B |
45.53%
|
01.02.2026 | Verified |
| Qwen3.5-27B |
41.87%
|
01.02.2026 | Unverified |
| Claude-Sonnet-4.5 |
52.44%
|
01.02.2026 | Unverified |
| Qwen3.5-35B-A3B |
38.41%
|
01.02.2026 | Unverified |
| Qwen3-235B-A22B |
5.08%
|
01.02.2026 | Unverified |
| Qwen3-VL-235B-A22B |
5.08%
|
01.02.2026 | Verified |
| GPT-5.4 |
80.08%
|
23.08.2026 | Verified |
| Claude Opus 4.6 |
75.41%
|
23.08.2026 | Verified |
| Kimi K2.5 |
56.30%
|
23.08.2026 | Verified |
| Kimi K3 |
100.00%
|
18.08.2026 | Verified |