Measured model

Measured on a Mac Studio (M5 Max, 36GB)DeepSeek-R1-Distill-Qwen-32B

DeepSeek · Ollama name hf.co/unsloth/DeepSeek-R1-Distill-Qwen-32B-GGUF:Q4_K_M

On a Mac Studio (M5 Max, 36GB), DeepSeek-R1-Distill-Qwen-32B answers a simple question in about 17.9 s, gets 81% of the Japanese knowledge questions right, 94% of the math problems, and uses 22.4GB of memory.

Results

MeasureResultAmong measured modelsNote
Reply timeabout 17.9 s31 of 33Until one simple common-sense question is answered (median of 50)
Knowledge (high school to university)81%12 of 33200 questions, ±6, 3 cut off (counted wrong)
Math (word problems)94%14 of 33100 questions, ±7
Memory used22.4GB32 of 33
EN→JA translation29 pts22 of 30out of 100 (chrF), about 18.4 s a sentence
JA→EN translation56 pts10 of 30out of 100 (chrF), about 16.7 s a sentence

"Among measured models" is the place among the 33 models measured on this site (lower is better for time and memory).

Try it on your Mac

  1. Install the Mac version of Ollama from ollama.com and start it.
  2. Type this in Terminal. The first time, the model downloads; then you can chat with it.ollama run hf.co/unsloth/DeepSeek-R1-Distill-Qwen-32B-GGUF:Q4_K_M

It used 22.4GB here. To use it comfortably alongside other apps, a Mac with 48GB of memory or more is a good guide (the model then takes about half the memory).

About the model

Maker
DeepSeek
Size
32.8B parameters
Quantization
Q4_K_M
Thinks before answering
Yes
Longest input at once
131,072 tokens
Released (Hugging Face)
Jan 20, 2025
Generation speed
20.5 tok/s
Start-up (loading)
about 1.6 s
Measured on
Oct 10, 2026
Ollama
0.34.3
Hugging Face
deepseek-ai/DeepSeek-R1-Distill-Qwen-32B
Series
DeepSeek

What the parts of a name mean (size, quantization and so on): how to read a model name.

Models of similar memory

Compare the main models on the text page. The series page lists every model measured.

Task recommendations and capabilities · Measurements by model