Measured model
Measured on a Mac Studio (M5 Max, 36GB)Ministral-3-3B-Instruct-2512
Mistral · Ollama name hf.co/mistralai/Ministral-3-3B-Instruct-2512-GGUF:Q4_K_M
On a Mac Studio (M5 Max, 36GB), Ministral-3-3B-Instruct-2512 answers a simple question in about 0.1 s, gets 48% of the Japanese knowledge questions right, 42% of the math problems, and uses 3.2GB 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 0.1 s | 1 of 30 | Until one simple common-sense question is answered (median of 50) |
| Knowledge (high school to university) | 48% | 30 of 30 | 200 questions, ±7, 2 cut off (counted wrong) |
| Math (word problems) | 42% | 30 of 30 | 100 questions, ±10, 4 cut off (counted wrong) |
| Memory used | 3.2GB | 4 of 30 | |
| EN→JA translation | 18 pts | 28 of 28 | out of 100 (chrF), about 0.3 s a sentence |
| JA→EN translation | 46 pts | 28 of 28 | out of 100 (chrF), about 0.2 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/mistralai/Ministral-3-3B-Instruct-2512-GGUF:Q4_K_M
It used 3.2GB here. To use it comfortably alongside other apps, a Mac with 8GB of memory or more is a good guide (the model then takes about half the memory).
About the model
- Maker
- Mistral
- Size
- 3.43B parameters
- Quantization
- Q4_K_M
- Thinks before answering
- No
- Longest input at once
- 262,144 tokens
- Released (Hugging Face)
- Oct 31, 2025
- Generation speed
- 141.8 tok/s
- Start-up (loading)
- about 0.8 s
- Measured on
- Sep 29, 2026
- Ollama
- 0.34.3
- Hugging Face
- mistralai/Ministral-3-3B-Instruct-2512
- Series
- Mistral
Models of similar memory
- NVIDIA-Nemotron-3-Nano-4B-BF16 (3.3GB of memory, knowledge 63%, reply about 2.4 s)
- Qwen3.5-4B (3.5GB of memory, knowledge 56%, reply about 40 s)
- MiniCPM5-2B (2GB of memory, knowledge 70%, reply about 1.5 s)
- LFM2.5-2.6B (2GB of memory, knowledge 73%, reply about 2.2 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.