NEW · API PLANS - 9 models, one key, 250M - 3B tokens a month · from /mo See plans →
Home › Models › Alibaba › Qwen3.5-397b-a17b

Qwen3.5-397b-a17b

Qwen3.5-397B-A17B is Alibaba's earlier sparse model, listed as coming soon. It is 397B total parameters with about 17B active, a 1M context window, and text plus image input. It is not the live Qwen 3.8 Flash, which is also 397B and already callable with an intelligence index of 28. No benchmark scores for this 3.5 checkpoint are published here.

Sparse architectureLong contextGeneral Reasoning
Parameters397B
Active params17B
Context1M