Q

qwen3.5-397b-a17b

Inndata:$75/M
Utdata:$300/M
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.
Ny
Kommersiell bruk