Mistral Small 4.
Mistral Small 4 is an Apache 2.0 mixture-of-experts model, 119B total and 6.5B active, with text and image input, 256k context and per-request reasoning.
2× Nvidia Q RTX 8000 · $0.54/hr.
Most-aggressive quantisation we have a working recommendation for. Lower precision = less VRAM = cheaper hardware, at a small accuracy cost.
Cheapest hosted endpoints.
Variants in the Mistral family.
Mistral's flagship open-weight model — 123B dense.
The current Mistral 7B — adds function calling + extended vocab.
Mistral 7B v0.2 — earlier 32K context revision.
Original Mistral 7B — historical reference.
Mistral × Nvidia collab — 12B Apache-licensed, multilingual.
Mistral Large 4 is a frontier multimodal (text and image input) model from Mi...
Mistral Large 3 2512 is Mistral’s most capable model to date, featuring a spa...
Mistral Medium 3 is a high-performance enterprise-grade language model design...
Mistral Medium 3.1 is an updated version of Mistral Medium 3, which is a high...
Mistral Medium 3.5 is a dense 128B open-weights model for instructions, reaso...
Mistral Small 3 is a 24B-parameter language model optimized for low-latency p...
Mistral Small 3.1 24B Instruct is an upgraded variant of Mistral Small 3 (250...
Mistral-Small-3.2-24B-Instruct-2506 is an updated 24B parameter model from Mi...
Mistral Saba is a 24B-parameter language model specifically designed for the ...
Frequently asked.
How do I run Mistral Small 4?
Where can I access Mistral Small 4?
How much does it cost to run Mistral Small 4?
Is Mistral Small 4 open-source or proprietary?
Cheapest hardware per quantisation.
Each row is one quantisation tier (the same weights compressed differently). Lower precision → lower VRAM → cheaper hardware, at the cost of small accuracy loss. $/hr is the cheapest rate on offer now. A multi-card setup uses a real instance of that size when one is rented; otherwise it is the single-card rate times the count.
| Quantisation | Cheapest GPU config | Total VRAM | Live $/hr | tokens/sec | |
|---|---|---|---|---|---|
|
FP16
FP16 — half precision (default)
|
384 GB |
$2.15/hr
est. 8 × single-card rate
Vast.ai · on-demand
|
— | Compare → | |
|
FP8
FP8 — 8-bit float (Hopper / Blackwell)
|
192 GB |
$1.08/hr
est. 4 × single-card rate
Vast.ai · on-demand
|
— | Compare → | |
|
INT4
INT4 — 4-bit integer (~4× VRAM saving)
|
96 GB |
$0.54/hr
est. 2 × single-card rate
Vast.ai · on-demand
|
— | Compare → |
What it costs per month across providers.
Estimate your monthly bill for Mistral Small 4 across every host that publishes per-token pricing. Slide your token volumes; the chart + table re-rank cheapest-first.
Cheapest provider on the left.
Total monthly cost — input + output tokens combined.
Bill breakdown.
Rent the GPU instead of paying per token.
For an open-weights model like Mistral Small 4, you can rent a GPU and serve inference yourself. The math: cheapest GPU rental × 730 hours/month. Power is included in a rental's hourly price.
Assumes the GPU runs 24/7 at ~85% utilisation. If your traffic is bursty, you'll pay less for the API and probably more for the GPU (idle hours still cost rental). The breakeven analysis lives on the Self-host vs API breakeven tool.
About Mistral Small 4.
Mistral Small 4 is an Apache 2.0 open-weights model that merges Mistral's Instruct, Reasoning and Devstral lines into one. It is a mixture-of-experts design with 119 billion parameters, 6.5 billion active per token, using 4 of 128 experts. It takes text and images, returns text, and has a 256k context. Reasoning is switched on or off per request. Mistral aims it at chat assistants, coding agents, document extraction and research. The maker also publishes an NVFP4 checkpoint and a speculative decoding head, and suggests its own API if local serving falls short.