qwen3.5-397b-a17b
1442 ? The Qwen3.5 series 397B-A17B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. It delivers...
LLM & Generative AI 1 model 5 listings snapshot Oct 14, 2026, 12:54 AM
Platforms
click a card to open the detail tableModel variants
W3r9 dedup — same model + identical tags merge; prices are listed per billing form| Model | Provider | Tags | Input /1M | Output /1M |
|---|---|---|---|---|
| qwen/qwen3.5-397b-a17b | OpenRouter | 397b256k contextmultimodalreasoning | $0.500 | $3.60 |
| Qwen/Qwen3.5-397B-A17B | DeepInfra | 397b256k contextmultimodalreasoning | $0.450 | $3.00 |
| Qwen/Qwen3.5-397B-A17B | Together | 397b256k context | $0.600 | $3.60 |
| qwen3.5-397b-a17b | DashScope | 397b | $0.178 | $1.07 |
| qwen/qwen3.5-397b-a17b | Novita | 397b | $0.600 | $3.60 |
Capabilities
official design_arena & architecture metadataDesign Arena Elo official design_arena benchmark (OpenRouter)
Input / Output modalities modality: text+image+video->text
Platform details
per provider| provider | model | Input /1M | Output /1M | Latency ? | Official $/1M | Availability | Data time |
|---|---|---|---|---|---|---|---|
| DashScope | qwen3.5-397b-a17b | $0.178 | $1.07 | — | — | — | 2026-10-14 00:53:13 |
| DeepInfra | Qwen/Qwen3.5-397B-A17B | $0.450 | $3.00 | — | — | — | 2026-10-14 00:43:44 |
| Novita | qwen/qwen3.5-397b-a17b | $0.600 | $3.60 | — | — | — | 2026-10-14 00:54:15 |
| OpenRouter | qwen/qwen3.5-397b-a17b | $0.500 | $3.60 | 128.1 ms | — | 100% available | 2026-10-14 00:43:21 |
| Together | Qwen/Qwen3.5-397B-A17B | $0.600 | $3.60 | — | — | — | 2026-10-14 00:53:13 |
5 rows · snapshot 2026-10-14 00:54:15 · from data/public/prices.json + measurements.json