| # | Model | Provider | Type | Score | |
|---|---|---|---|---|---|
| 1 | Qwen3.7 Plus | Alibaba | Proprietary | 81.7% | |
| 2 | Step 3.7 Flash | StepFun | Open weights | 79.2% | |
| 3 | Qwen3.8 Max | Alibaba | Open weights | 75.0% | |
| 4 | dots3-note Preview | Dots Studio · Open weight | 72.5% | ||
| 5 | Gemini 3.1 Pro | Google · Closed | 72.4% | ||
| 6 | Muse Spark | Meta · Closed | 71.3% | ||
| 7 | GPT-5.4 | OpenAI · Closed | 61.1% | ||
| 8 | Qwen3.6-35B-A3B | Alibaba · Open weight | 58.9% | ||
| 9 | Grok 4.20 | xAI · Closed | 57.4% | ||
| 10 | Qwen3.6-27B | Alibaba · Open weight | 56.1% | ||
| 11 | LFM2.5-VL-3B | LiquidAI · Open weight | 35.4% |