Qwen3-Next-80B-A3B-Instruct is the next generation basic model released by Alibaba Tongyi Qianwen team. It is based on the new Qwen3-Next architecture and is designed for ultimate training and inference efficiency. The model uses an innovative hybrid attention mechanism (Gated DeltaNet and Gated Attention), a high sparsity mixed expert (MoE) structure and multiple training stability...
Provider: Qwen
Context window: 256K
Pricing: $0.06 input / $0.55 output per 1M tokens
Capabilities: serverless, json_mode, function_calling, structured_output
