Qwen3-NextOpen Weight
Released
Qwen3-Next is an AI model released by Qwen on Thursday, Sep 11 2025, 51 days after Qwen3-Coder. It is an open-weight model — the trained weights are available to download and run. It is a 80B parameterThe learned values that shape how a model responds. B means billion and T means trillion. For mixture-of-experts models, the total count is larger than the number active for each token. Parameter count alone does not measure quality. model with a 256k token context windowHow much information the model can work with in one request, measured in tokens — pieces of words or other data. This includes your question, attachments, conversation and reply. Limits can vary by model variant and provider..
Get Qwen releases and AI news.
Every other Monday
Available from
| DeepInfra | flex | $0.072 | $0.88 | 256K | fp8 |
|---|---|---|---|---|---|
| DeepInfra | — | $0.09 | $1.10 | 256K | fp8 |
| Parasail | — | $0.10 | $1.10 | 256K | fp8 |
| DeepInfra | priority | $0.135 | $1.65 | 256K | fp8 |
Compare Qwen3-Next with
Suggested comparisons
Qwen3-NextvsQwen3.8-Max-0902Qwen3-NextvsGPT-6.1 TerraQwen3-NextvsClaude Haiku 5.5Qwen3-NextvsGemini 4 CarbonQwen3-NextvsMuse Spark OpenQwen3-NextvsGrok 4.7Qwen3-NextvsDeepSeek-V4.1-Flash-1017Qwen3-NextvsMistral Large 4Qwen3-NextvsKimi K3.1 ProQwen3-NextvsGLM-5.3-FlashQwen3-NextvsNemotron 3.5 Bolt E4B 4GBAbout
Qwen3-Next introduced a more efficient model architecture in September 2025, beginning with an 80-billion-parameter model that activated about 3 billion parameters per token. It combined Gated DeltaNet with gated attention and a highly sparse mixture-of-experts design. The objective was to increase capacity and handle long inputs without proportionately increasing the computation required for every generated token.
Alibaba published separate Instruct and Thinking checkpoints, with the Instruct version operating without explicit thinking blocks. The documented native context was 262,144 tokens, with extensions requiring the prescribed configuration. This made Qwen3-Next an architectural development within the Qwen line, rather than simply a larger dense model. Its overall weight size and active parameter count described different aspects of deployment cost.
Frequently asked questions
Qwen3-Next was released by Qwen on Thursday, Sep 11 2025.