VentureBeat
Qwen3-Coder-Next offers vibe coders a powerful open source, ultra-sparse model with 10x higher throughput for repo tasks

Alibaba's Qwen team released Qwen3-Coder-Next, a 80-billion-parameter model with ultra-sparse architecture. It delivers elite performance with low deployment costs and high throughput. The model supports 370 programming languages and outperforms larger models in benchmark evaluations.