by AlibabaRank #66Score 75.6
The Qwen3.5 Series 35B-A3B is a native vision-language model designed with a hybrid architecture that integrates linear attention mechanisms and a sparse mixture-of-experts model, achieving higher inference efficiency. Its overall...
| Signal | Normalized | Weight | Contribution | Freshness |
|---|---|---|---|---|
Capabilities capability | 83.3 | 20% | 16.7 | 2026-06-24T09:59:39.964Z |
Benchmarks benchmark | 75.6 | 30% | 22.7 | 2026-06-24T09:59:39.964Z |
Pricing pricing_tier | 99.0 | 15% | 14.9 | 2026-06-24T09:59:39.964Z |
Context Window context_window | 77.4 | 10% | 7.7 | 2026-06-24T09:59:39.964Z |
Recency recency | 100.0 | 15% | 15.0 | 2026-06-24T09:59:39.964Z |
Output Capacity output_capacity | 90.3 | 10% | 9.0 | 2026-06-24T09:59:39.964Z |
| Capability | Supported |
|---|---|
| Vision | Yes |
| Reasoning | Yes |
| JSON Mode | Yes |
| Streaming | Yes |
| Function Calling | Yes |
| Web Search | No |