Magistral Small — local AI model by Mistral AI
Written by Jakub Rusinowski · Last updated
Mistral's reasoning line: Mistral Small with long chain-of-thought added via SFT on Magistral Medium traces plus RL. Apache 2.0, vision-capable, and sized so the whole thing fits a single RTX 4090 or a 32 GB MacBook once quantized — which is the point. Reasoning traces degrade past roughly 40K tokens even though the context window runs to 128K.
Variants
The smallest Magistral Small variant needs about 15 GB of VRAM at Q4_K_M — quantized weights plus framework overhead, before any KV cache.
| Model | VRAM at Q4 | VRAM | Context | Run it |
|---|---|---|---|---|
| Magistral Small 24B → 24B | ~15.3 GB | 131,072 | ollama run magistral:24b |
Memory is quantized weights plus overhead at Q4_K_M, from the same engine as the GPU & VRAM checker.
How to run Magistral Small locally
Install Ollama, then pull the tag.
ollama run magistral:24bPick a size above for its own VRAM figure, speed estimate and install command.
Licence
Commercial use permitted. No usage restrictions beyond attribution.
Applies to: Magistral Small 24BRecommended GPU
The cheapest catalogued GPU that runs Magistral Small locally (min 15 GB VRAM) is the AMD Radeon RX 7900 XT (20 GB).
Can I run Magistral Small on my GPU?
- Magistral Small on AMD Radeon RX 7800 XT
- Magistral Small on AMD Radeon RX 7900 XT
- Magistral Small on AMD Radeon RX 7900 XTX
- Magistral Small on AMD Radeon RX 9060 XT 16GB
- Magistral Small on AMD Radeon RX 9060 XT 8GB
- Magistral Small on AMD Radeon RX 9070
- Magistral Small on AMD Radeon RX 9070 XT
- Magistral Small on Apple M1
- Magistral Small on Apple M1 Pro
- Magistral Small on Apple M2
- Magistral Small on Apple M2 Pro
- Magistral Small on Apple M3
- Magistral Small on Apple M3 Pro
- Magistral Small on Apple M4
- Magistral Small on Apple M4 Pro
- Magistral Small on Apple M5
- Magistral Small on Intel Arc B570
- Magistral Small on Intel Arc B580
- Magistral Small on NVIDIA GeForce RTX 3060 (12GB)
- Magistral Small on NVIDIA GeForce RTX 3070
- Magistral Small on NVIDIA GeForce RTX 3070 Ti
- Magistral Small on NVIDIA GeForce RTX 3080 (10GB)
- Magistral Small on NVIDIA GeForce RTX 3090
- Magistral Small on NVIDIA GeForce RTX 4060
- Magistral Small on NVIDIA GeForce RTX 4060 Ti 16GB
- Magistral Small on NVIDIA GeForce RTX 4070
- Magistral Small on NVIDIA GeForce RTX 4070 Super
- Magistral Small on NVIDIA GeForce RTX 4070 Ti
- Magistral Small on NVIDIA GeForce RTX 4070 Ti Super
- Magistral Small on NVIDIA GeForce RTX 4080
- Magistral Small on NVIDIA GeForce RTX 4080 Super
- Magistral Small on NVIDIA GeForce RTX 4090
- Magistral Small on NVIDIA GeForce RTX 5060
- Magistral Small on NVIDIA GeForce RTX 5060 Ti 16GB
- Magistral Small on NVIDIA GeForce RTX 5060 Ti 8GB
- Magistral Small on NVIDIA GeForce RTX 5070
- Magistral Small on NVIDIA GeForce RTX 5070 Ti
- Magistral Small on NVIDIA GeForce RTX 5080
- Magistral Small on NVIDIA GeForce RTX 5090
Magistral Small — frequently asked questions
How much VRAM does Magistral Small need?
Magistral Small needs about 15 GB VRAM at Q4_K_M quantization for its smallest variant. Variants: Magistral Small 24B (15 GB, Q4_K_M). On Apple Silicon, unified memory counts toward this requirement.
Can I run Magistral Small on an RTX 4090 (24 GB)?
Yes — Magistral Small runs on an RTX 4090 (24 GB) and other 24 GB cards such as the RTX 3090. Smaller variants also fit comfortably on 8–16 GB GPUs at Q4_K_M.
What quantization should I use for Magistral Small?
Q4_K_M is the best balance of quality and VRAM for Magistral Small in most cases. Choose Q8_0 for near-lossless quality if you have spare VRAM, or smaller quants (Q3/Q2) only when memory is tight.
How do I run Magistral Small with Ollama?
Install Ollama, then run: ollama run magistral:24b. This downloads Magistral Small and starts a local, OpenAI-compatible endpoint — no internet connection is needed after the initial download.