DeepSeek R1 Distill Qwen 32B

DeepSeek · reasoning · 80K context · 4× credits

Reasoning model distilled from DeepSeek R1. Shows its thinking; best for math and logic.

Model ID: mixerlead-ai/deepseek-r1-distill-qwen-32b

Best for

A reasoning model distilled from DeepSeek R1. It writes out its thinking before the answer.

Maths and logic where you want to see the working

Teaching or tutoring apps

The reply contains its thinking inside <think>...</think>. Remove that part before showing the answer to users.

Thinking uses many tokens and this model costs 4×. For cheaper reasoning, try GPT-OSS 120B (0.75×).

Set max_tokens to 1500 or more.