Alibaba / Qwen

Alibaba's 397B MoE open-weight flagship with 17B active parameters and native multimodal support

Alibaba's Qwen3 Coder Next, a coding-specialized model for local development built on the Qwen3 Next architecture

Alibaba's Qwen3-VL Flash, a faster multimodal variant from the Qwen3-VL family for efficient vision tasks

Alibaba's Qwen3-VL-235B-A22B multimodal MoE model with 235B parameters and 256K context

Alibaba's Qwen3-Max, a proprietary trillion-parameter model with 262K context and hybrid reasoning

Alibaba's Qwen3-Coder-480B MoE model with 480B parameters, 256K context, and agentic coding capabilities

Alibaba's Qwen3-235B-A22B MoE model with 235B parameters, 256K context, and hybrid thinking

Alibaba's compact multimodal model supporting text, image, audio, and video inputs and outputs

Alibaba's 32B vision-language model for document understanding, chart analysis, and visual reasoning

Alibaba's code-focused 7B model specialized for programming tasks across 92 languages