Home
LLMs
Mistral Small 4

Mistral Small 4

A 119B open-weight MoE under Apache 2.0 with 128 experts and roughly 6B active parameters per token, unifying Mistral Small, Magistral and Devstral into one hybrid model.

Params: 119B, 6.5B active
Context: 256K tokens
Released: Mar 2026
License: Apache 2.0

Visit Mistral Compare vs other LLMs →

Hosted API price

List price via Mistral · per 1M tokens

USD

Input: $0.15 /1M
Output: $0.60 /1M
Hosted by: 2 providers

Hosted API pricing

How we price

Order

Providers are listed cheapest input rate first, then by output rate, a tie going to the creator's own listing. A provider that lists the model without a published rate sits at the end.

Prices

Base-tier, on-demand rates in USD per 1M tokens, converted at ECB reference rates where a provider publishes in another currency. Cached-input and batch rates show only where a provider publishes them. "Cheapest" marks the one provider with the lowest input and output pair; a tie wears no badge.

Cost

Input rate times the input tokens plus output rate times the output tokens, at the monthly volume set above the table. Cached-input, batch, long-context and reasoning-token billing are not modelled.

Transparency and funding

Affiliates: Affiliate links are marked. We may earn a commission if you click them, but commissions never affect the order.

Every provider serving Mistral Small 4, cheapest input first. Set your monthly volume to see what each would bill.

Input M Output M

Every provider serving Mistral Small 4, with its price per 1M input and output tokens and what the volume above would cost
Provider	Input /1M	Output /1M	Cost at 10M in + 2M out	Link
Mistral Creator	$0.15	$0.60	$2.70	Visit website
EmpirioLabs AI	$0.15	$0.60	$2.70	Visit website

Heads up: Base-tier, on-demand rates per 1M tokens; cached, batch and long-context tiers excluded. A provider may serve a shorter context or a quantized build than the creator's release. Verify before provisioning. More on how we price.

Estimated cost to self-host

Mistral Small 4 needs about 76 GB of GPU memory at 4-bit with 32K context. The cheapest rental that fits is 8× RTX 3060 at about $403 a month.

That costs the same as roughly 2B tokens a month on Mistral's API. Self-hosting is more expensive below that volume.

Single node ≤ 8

What it costs to self-host Mistral Small 4 at each precision: the memory it needs, the cheapest rental that holds it, that rental's month and the API volume it pays off at
Precision	Memory	Cheapest fit (32K context)(Full 256K context)	Cost /mo	Break-even vs API
4-bit INT4 / FP4	76 GB100 GB	8× RTX 3060	$403$691	2B tokens /mo3B tokens /mo
8-bit FP8 / INT8	129 GB153 GB	8× RTX 3090	$1,267$2,534	6B tokens /mo11B tokens /mo
16-bit FP16 / BF16	248 GB304 GB	8× A40	$2,822$2,822	13B tokens /mo13B tokens /mo

Estimates based on median on-demand rates for Nvidia GPUs. Memory is weights plus KV cache for one request, using FP8 KV cache where supported and FP16/BF16 otherwise. KV cache assumed at 250 KB per token, a conservative default. Break-even assumes a 5:1 input-to-output ratio. No guarantee of runtime support, usable performance, or that a matching quantized build exists. How we estimate costs.

Similarly priced models

The models nearest Mistral Small 4 by blended rate, each at its own cheapest provider.

Models priced nearest Mistral Small 4 per 1M tokens, each at its cheapest provider
Model	Blended /1M	Input /1M	Output /1M	Context	Cutoff	vs Mistral Small 4
Qwen3-VL-8B Alibaba Cloud	$0.2083	$0.15	$0.50	262K		−7%
Qwen-MT-Turbo Alibaba Cloud	$0.215	$0.16	$0.49	16K		−4%
Command R 08-2024 Cohere	$0.225	$0.15	$0.60	128K	Jun 2024	0%
GPT-4o Mini OpenAI	$0.225	$0.15	$0.60	128K	Oct 2023	0%
MiniMax-M2.7 MiniMax	$0.225	$0.15	$0.60	205K		0%
Mistral Small 4 This model Mistral	$0.225	$0.15	$0.60	256K
Qwen3.8-27B Alibaba Cloud	$0.225	$0.17	$0.50	262K		0%
Qwen3-Coder-Next Alibaba Cloud	$0.23	$0.08	$0.98	262K		+2%
Llama 4 Scout Meta	$0.24	$0.16	$0.64	10M	Aug 2024	+7%
Qwen3.5-122B-A10B Alibaba Cloud	$0.2533	$0.12	$0.92	262K		+13%
Llama 4 Maverick Meta	$0.2833	$0.20	$0.70	1M	Aug 2024	+26%

USD per 1M tokens at each model's cheapest listed provider. Blended is the cost of 10M input plus 2M output tokens, spread over the 12M.

Vast.ai

Per-second billing. Pay only for what you use.

50+ GPU models from RTX 3060 to B200, from $0.04/hr

No minimums, no long-term contracts.

SOC2 certified · PyTorch, CUDA & TensorFlow templates ready

Learn more

Capabilities

What Mistral Small 4 accepts and can do, as published by Mistral.

Text Accepts and generates natural-language text.
Image Accepts images as input alongside text.
Reasoning Thinks before it answers, always on or as a switchable mode.
Function calling Connect to external tools, APIs, and systems.
Structured output Return responses in structured formats like JSON.
Web search Search the web for up-to-date information.
Code execution Write and execute code in a sandboxed environment.
File upload Process uploaded documents, spreadsheets, and other files.
Image generation Generate and edit images from text prompts.

Common questions

What is Mistral Small 4 good for?

Instruct, reasoning and coding in one model, which Mistral says unifies Mistral Small, Magistral and Devstral. Apache 2.0 weights, servable on a single high-memory GPU.

When is Mistral Small 4 not a good fit?

Self-hosting on a single consumer GPU: only 6.5B parameters are active per token, but all 119B must be loaded. No audio or video input, and no maximum output length is published.

What is the cheapest way to run Mistral Small 4?

Hosted, unless you push serious volume. Mistral charges $0.15 in / $0.60 out per 1M tokens. The cheapest rental that fits is 8x RTX 3060 at $403 a month, which costs the same as about 2B tokens a month on that API.

How does Mistral Small 4 compare with Mistral Small 3.2?

Mistral says Mistral Small 4 is the first Mistral model to fold Magistral reasoning, Pixtral vision and Devstral coding into one mixture-of-experts model, replacing the dense Mistral Small 3.2, with a reasoning effort setting and faster serving. Same text and image input. Mistral Small 4 came out 9 months later and has a larger context (256K against 128K tokens) and more parameters (119B with 6.5B active against 24B).

Can I self-host Mistral Small 4?

Yes. The weights are Apache 2.0 licensed. At 4-bit it needs about 76 GB of GPU memory, which starts at roughly $403 a month on the cheapest rental that fits.