Qwen 3 235B A22B Instruct 2507 TPUT

Cost-optimized variant of the 235B model for high-throughput general instruction tasks.

A high-throughput instruct variant of the 235B family, targeting general instruction-following across a broad range of tasks. It keeps the strong reasoning and alignment characteristics of the 2507 release while tuning for production-serving economics. The model works well as a high-end general assistant in enterprises, where it can power everything from code copilots to knowledge workers. It is particularly valuable when you need one powerful model to cover many use cases efficiently.

image of technology-driven data analysis scene

Explore other AI Models