Magistral-Small-2509 measured on a Mac Studio (M5 Max, 36GB)

Results

Measure	Result	Among measured models	Note
Reply time	about 28.8 s	31 of 33	Until one simple common-sense question is answered (median of 50)
Knowledge (high school to university)	28%	33 of 33	200 questions, ±7
Math (word problems)	1%	33 of 33	100 questions, ±4
Memory used	16.2GB	19 of 33
EN→JA translation	0 pts	30 of 30	out of 100 (chrF), about 29 s a sentence
JA→EN translation	6 pts	30 of 30	out of 100 (chrF), about 28.9 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

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/Magistral-Small-2509-GGUF:Q4_K_M

It used 16.2GB here. To use it comfortably alongside other apps, a Mac with 36GB of memory or more is a good guide (the model then takes about half the memory).

About the model

Maker: Mistral
Size: 23.6B parameters
Quantization: Q4_K_M
Thinks before answering: No
Longest input at once: 40,960 tokens
Released (Hugging Face): Sep 12, 2025
Generation speed: 28.9 tok/s
Start-up (loading): about 1.3 s
Measured on: Oct 3, 2026
Ollama: 0.34.3
Hugging Face: mistralai/Magistral-Small-2509
Series: Mistral

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

Models of similar memory

Devstral-Small-2-24B-Instruct-2512 (16.3GB of memory, knowledge 66%, reply about 0.2 s)
Mistral-Small-3.1-24B-Instruct-2503 (16.3GB of memory, knowledge 72%, reply about 0.2 s)
Mistral-Small-3.2-24B-Instruct-2506 (16.3GB of memory, knowledge 77%, reply about 0.2 s)
muse-glimmer:30b (17.6GB of memory, knowledge 83%, reply about 14.3 s)
LFM2-24B-A2B (14.8GB of memory, knowledge 54%, reply about 0.1 s)

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