All AI models · LLM benchmarks · Methodology

DeepSeek

DeepSeek R1 Distill Qwen 32B

Classic open reasoning model for transparent math, coding, and deliberate problem solving

Model facts

Context window
8K tokens
Maximum output
8K tokens
Cheapest paid input
$0.3 per 1M tokens
Cheapest paid output
$0.3 per 1M tokens
Open weights
yes
Providers
3

Capabilities and modalities

reasoning, tool calling, temperature control, open weights, text.

Local hardware estimate

At 4K context: Q4_K_M 21.0 GiB, Q8_0 36.3 GiB, F16 64.9 GiB working memory.

Parameters
32.8 billion
Architecture
qwen2
Model license
mit

Planning estimate adapted from the MIT-licensed whichllm estimator using metadata from Hugging Face; not a vendor minimum requirement.

Artificial Analysis benchmarks

Compare this model on the full LLM benchmark leaderboard.

ConfigurationAAICodingMathOutput tokens/s
default11.0not measured63.0not measured

Observed price history

  • 2026-08-24: $0.3 input / $0.3 output per 1M tokens

API providers

  • NovitaAI — model id deepseek/deepseek-r1-distill-qwen-32b; $0.3 input / $0.3 output per 1M tokens; provider documentation
  • Alibaba (China) — model id deepseek-r1-distill-qwen-32b; $0.287 input / $0.861 output per 1M tokens; provider documentation
  • Cloudflare Workers AI — model id @cf/deepseek-ai/deepseek-r1-distill-qwen-32b; $0.497 input / $4.88 output per 1M tokens; provider documentation