Alibaba Qwen models

Alibaba publishes Qwen models available through ApiFlux.

Qwen3.8 Max

Alibaba's flagship 2.4-trillion-parameter Qwen mixture-of-experts model with a 1M-token context, multimodal input, and controllable reasoning.

text

A cost-effective Qwen3.7 model for general-purpose, coding, and agentic workloads.

text

Alibaba's flagship Qwen3.7 model for coding, productivity, and agent-centric workloads.

text

An open-weight Qwen mixture-of-experts model designed for efficient multimodal reasoning.

text

A dense 27-billion-parameter Qwen model for general-purpose and multimodal tasks.

text

A scalable Qwen model combining efficient attention with sparse expert routing for high-performance inference.

text

An efficient Qwen Flash model combining linear attention with sparse mixture-of-experts routing.

text

An instruction-tuned Qwen chat model optimized for fast, stable responses across reasoning, coding, and multilingual tasks.

text

A multilingual Qwen mixture-of-experts model tuned for instruction following and text generation.

text

Enterprise AI Gateway for routing, securing, and observing every model call.