Qwen: Qwen3 Next 80B A3B Thinking

Qwen3-Next-80B-A3B-Thinking is a cutting-edge, reasoning-first chat model from the Qwen3-Next series. It's specifically engineered to tackle challenging multi-step problems, including intricate math proofs, sophisticated code synthesis and debugging, complex logic puzzles, and advanced agentic planning. This model distinguishes itself by outputting structured “thinking” traces by default, providing transparency and clarity in its problem-solving process. It reports strong results across knowledge, reasoning, coding, alignment, and multilingual evaluations, emphasizing stability under long chains of thought and efficient scaling during inference. This model is highly suitable for integration into agent frameworks and tool use (function calling), retrieval-heavy workflows, and standardized benchmarking where detailed, step-by-step solutions are crucial. It supports long, detailed completions and leverages throughput-oriented techniques like multi-token prediction for faster generation. With a generous context window of 131K tokens and a max output of 4K tokens, it handles extensive interactions. Pricing is competitive at $0.15/$1.20 per 1M input/output tokens, available on our PRO access tier. Note that it operates in thinking-only mode, focusing purely on its reasoning capabilities.