Neura Intelligence Index
Back to Leaderboard
OSWorld
Real-world computer task completion in operating system environments.
agent0-100 percent
Model Rankings (20)
| # | Model | Score | Provider | Source |
|---|---|---|---|---|
| 1 | B Seed 2.1 Pro🇨🇳 | 78.8 | ByteDance | verified |
| 2 | B Seed 2.1 Turbo🇨🇳 | 76.4 | ByteDance | verified |
| 3 | 72.7 | Anthropic | verified | |
| 4 | 72.5 | Anthropic | verified | |
| 5 | Q Qwen3 VL 235B A22B Instruct🇨🇳 | 66.7 | Alibaba Cloud / Qwen Team | verified |
| 6 | 66.3 | Anthropic | verified | |
| 7 | Z GLM-5V-Turbo🇨🇳 | 62.3 | Zhipu AI | verified |
| 8 | 61.4 | Anthropic | verified | |
| 9 | 50.7 | Anthropic | verified | |
| 10 | Q Qwen3 VL 32B Thinking🇨🇳 | 41 | Alibaba Cloud / Qwen Team | verified |
| 11 | Q Qwen3 VL 235B A22B Thinking🇨🇳 | 38.1 | Alibaba Cloud / Qwen Team | verified |
| 12 | Q Qwen3 VL 8B Thinking🇨🇳 | 33.9 | Alibaba Cloud / Qwen Team | verified |
| 13 | Q Qwen3 VL 8B Instruct🇨🇳 | 33.9 | Alibaba Cloud / Qwen Team | verified |
| 14 | Q Qwen3 VL 32B Instruct🇨🇳 | 32.6 | Alibaba Cloud / Qwen Team | verified |
| 15 | Q Qwen3 VL 4B Thinking🇨🇳 | 31.4 | Alibaba Cloud / Qwen Team | verified |
| 16 | Q Qwen3 VL 30B A3B Thinking🇨🇳 | 30.6 | Alibaba Cloud / Qwen Team | verified |
| 17 | Q Qwen3 VL 30B A3B Instruct🇨🇳 | 30.3 | Alibaba Cloud / Qwen Team | verified |
| 18 | Q Qwen3 VL 4B Instruct🇨🇳 | 26.2 | Alibaba Cloud / Qwen Team | verified |
| 19 | Q Qwen2.5 VL 72B Instruct🇨🇳 | 8.8 | Alibaba Cloud / Qwen Team | verified |
| 20 | Q Qwen2.5 VL 32B Instruct🇨🇳 | 5.9 | Alibaba Cloud / Qwen Team | verified |