Skip to main content
Language modelActive

Qwen 3 Next 80B A3B Instruct

Alibaba

Released
September 9, 2025
Data date
September 8, 2026
Speed
177.7 tokens/s
Output throughput
Price
$0.1 · $1.1
Input · Output / 1M tokens
Input
Text
Output
Text

Qwen 3 Next 80B A3B Instruct separates direct answers from the Thinking checkpoint built on the same 80B-A3B foundation. Three billion of its 80 billion parameters are active per token; the instruct variant accepts text and runs in non-thinking mode. The native window is 262,144 tokens, with extension to 1,010,000 according to the model card.

For visible reasoning, use the equally sized Qwen 3 Next 80B A3B Thinking checkpoint. They share the MoE foundation and use Apache-2.0.

Specifications and access

SpecificationValue and source
Weights
Open weightsSource
Access
API, LocalSource
Parameters
80BSource
Active parameters
3BSource
Context window
262,144 tokensSource
Weights repository
Qwen/Qwen3-Next-80B-A3B-InstructSource
License
Apache-2.0Source
Architecture
Qwen3 causal language model with sparse MoE and 8 activated expertsSource
Input
TextSource
Output
TextSource
Reasoning mode
non-thinking onlySource
Tool calling
YesSource
Structured output
YesSource
Native context window
262,144 tokensSource
Maximum context window
1,010,000 tokensSource
Parameters
80 BSource
Active parameters
3 BSource
Languages
119 languages and dialectsSource
Repository created
2025-09-09 (Repository creation date is a precise publication proxy, but may differ from the public announcement date or an individual checkpoint's first availability.)Source
Training tokens
36 trillionSource