Qwen
Qwen: Qwen3 Next 80B A3B Thinking
Qwen3-Next-80B-A3B-Thinking is a reasoning-first chat model in the Qwen3-Next line that outputs structured “thinking” traces by default. It’s designed for hard multi-step problems; math proofs, code synthesis/debugging, logic, and agentic...
| Specifications | |
|---|---|
| Provider | Qwen |
| Input pricing | $0.10/M tokens |
| Output pricing | $0.78/M tokens |
| Context length | 262K tokens |
| Max output | 33K tokens |
| Input modalities | text |
| Output modalities | text |
| Tokenizer | Qwen3 |
| Knowledge cutoff | 2025-09-30 |
| Content moderated | No |
| HuggingFace | Qwen/Qwen3-Next-80B-A3B-Thinking |
Supported Parameters
frequency_penaltyinclude_reasoninglogprobsmax_tokenspresence_penaltyreasoningrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_logprobstop_p