Measured model
Measured on a Mac Studio (M5 Max, 36GB)DeepSeek-R1-0528-Qwen3-8B
DeepSeek · Ollama name hf.co/unsloth/DeepSeek-R1-0528-Qwen3-8B-GGUF:Q4_K_M
On a Mac Studio (M5 Max, 36GB), DeepSeek-R1-0528-Qwen3-8B answers a simple question in about 9.8 s, gets 73% of the Japanese knowledge questions right, 78% of the math problems, and uses 6.4GB of memory.
These numbers are from before the method changed on Oct 2, 2026 (see how it was measured).
Results
| Measure | Result | Among measured models | Note |
|---|---|---|---|
| Reply time | about 9.8 s | 26 of 30 | Until one simple common-sense question is answered (median of 50) |
| Knowledge (high school to university) | 73% | 17 of 30 | 200 questions, ±7, 15 cut off (counted wrong) |
| Math (word problems) | 78% | 21 of 30 | 100 questions, ±9, 7 cut off (counted wrong) |
| Memory used | 6.4GB | 11 of 30 | |
| EN→JA translation | 25 pts | 26 of 28 | out of 100 (chrF), about 6.2 s a sentence |
| JA→EN translation | 51 pts | 24 of 28 | out of 100 (chrF), about 5.8 s a sentence |
"Among measured models" is the place among the 30 models measured on this site (lower is better for time and memory).
Try it on your Mac
- Install the Mac version of Ollama from ollama.com and start it.
- Type this in Terminal. The first time, the model downloads; then you can chat with it.
ollama run hf.co/unsloth/DeepSeek-R1-0528-Qwen3-8B-GGUF:Q4_K_M
It used 6.4GB here. To use it comfortably alongside other apps, a Mac with 16GB of memory or more is a good guide (the model then takes about half the memory).
About the model
- Maker
- DeepSeek
- Size
- 8.19B parameters
- Quantization
- Q4_K_M
- Thinks before answering
- Yes
- Longest input at once
- 131,072 tokens
- Released (Hugging Face)
- May 29, 2025
- Generation speed
- 77.3 tok/s
- Start-up (loading)
- about 0.6 s
- Measured on
- Sep 30, 2026
- Ollama
- 0.34.3
- Hugging Face
- deepseek-ai/DeepSeek-R1-0528-Qwen3-8B
- Series
- DeepSeek
Models of similar memory
- Ornith-1.5-9B (6.4GB of memory, knowledge 78%, reply about 2.5 s)
- Ornith-1.0-9B (6.2GB of memory, knowledge 83%, reply about 3.6 s)
- GLM-4-9B-0414 (6.7GB of memory, knowledge 66%, reply about 0.1 s)
- gemma-4-E4B-it (5.4GB of memory, knowledge 82%, reply about 4 s)
- LFM2.5-8B-A1B (5.4GB of memory, knowledge 67%, reply about 1.5 s)
Compare the main models on the text page. The series page lists every model measured.