Skip to content

Qwen3MoE

This model was released on 2025-04-29 and added to Hugging Face Transformers on 2025-03-31.

Qwen3MoE refers to the mixture of experts model architecture Qwen3-235B-A22B which was released with its dense variant Qwen3 (blog post).

To be released with the official model launch.

To be released with the official model launch.

[[autodoc]] Qwen3MoeConfig

[[autodoc]] Qwen3MoeModel - forward

[[autodoc]] Qwen3MoeForCausalLM - forward

[[autodoc]] Qwen3MoeForSequenceClassification - forward

[[autodoc]] Qwen3MoeForTokenClassification - forward

[[autodoc]] Qwen3MoeForQuestionAnswering - forward