Catalog
Mistralrouted

Mistral Small 24B (2501)

Mistral Small 24B (release 2501, marketed as Mistral Small 3) is Mistral AI's dense 24B instruction model under Apache 2.0. Its makers built it for fast conversational agents and low-latency function calling, and list dozens of supported languages, among them English, French, German, Spanish, Italian, Portuguese and Polish. No thinking mode and a short context window by current standards — a quick, predictable model for short exchanges from a European developer. It is fulfilled through global providers under our European contract, and requests go only to backends that do not retain prompts or completions.

model id: mistral-small-24b

from openai import OpenAI client = OpenAI(    api_key="hb-...",    base_url="https://api.heabsy.com/v1",) r = client.chat.completions.create(    model="mistral-small-24b",    messages=[{"role": "user", "content": "Hello"}],)print(r.choices[0].message.content)
Context
32K tokens
Serving
Global providers · EU contract
Tool calling
Reasoning
Structured output
Vision
Input
$0.05/ 1M
Output
$0.08/ 1M

Price list, updated 12 September 2026. No minimum spend, no subscription.

About the model

Developer
Mistral AI
Released
January 2025
Licence
Apache-2.0
Architecture
dense · 24B parameters
Input
text
Reasoning
no reasoning mode
Languages named by the developer
English, French, German, Spanish, Italian, Portuguese, Dutch, Polish, Russian, Chinese, Japanese, Korean

Source: the developer's model card, as of 17 September 2026.

Licence: what it means for your company

Apache-2.0. A permissive open-source licence: you may use, modify and sell products built on the model, including commercially; when you redistribute the weights, the licence and copyright notices go with them.

A summary of the licence, not legal advice.

Benchmarks reported by the developer

MMLU-Pro
66.3 %
HumanEval
84.8 %
IFEval
82.9 %

Figures from the developer's model card, not our measurement. Source: huggingface.co/mistralai/Mistral-Small-24B-Instruct-2501

Settings recommended by the developer

  • temperature 0.15

Source: huggingface.co/mistralai/Mistral-Small-24B-Instruct-2501

Technical details

Context window
32K tokens
Max output
16,384 tokens
Throughput limit
200K tok/min
Tokenizer
Other
Streaming
yes
Open weights
mistralai/Mistral-Small-24B-Instruct-2501
Supported parameters
max_tokens
1 … 16384
temperature
0 … 2
top_p
0 … 1
frequency_penalty
-2 … 2
presence_penalty
-2 … 2
stop
array
seed
integer
Bought for you, billed by us

This model is fulfilled through global providers under our European contract: one DPA with an EU company, one invoice for every model in the catalog, no US counterparty on your paperwork. Compute may run outside the EEA — if you need strict EEA-only processing, use our own-hardware model.

Questions about this model

Where is Mistral Small 24B computed?

Outside the EEA, through a global provider under our European contract. You get one DPA with an EU company and one invoice, but the compute itself does not run in the EEA — we state that instead of hiding it.

Does Mistral Small 24B store prompts and completions?

No. This model runs with zero data retention: request content is not written to durable storage and nothing is used for training.

How large is the Mistral Small 24B context window?

32K tokens, with up to 16,384 tokens of output.

What does Mistral Small 24B cost?

$0.05 per million input tokens and $0.08 per million output tokens. No minimum spend and no subscription; the same price list covers the API and the invoice.

Is there a throughput limit on Mistral Small 24B?

Yes, 200K tokens per minute. Higher limits are a question of volume, not of principle.

Which languages does Mistral Small 24B support?

The developer names English, French, German, Spanish, Italian, Portuguese, Dutch, Polish, Russian, Chinese, Japanese, Korean. Languages not on that list may still work, but they are not claimed — test on your own material first.

Under which licence is Mistral Small 24B released?

Apache-2.0. A permissive open-source licence: you may use, modify and sell products built on the model, including commercially; when you redistribute the weights, the licence and copyright notices go with them.

Can reasoning be switched off in Mistral Small 24B?

Mistral Small 24B has no reasoning mode — it answers directly, which keeps latency and output tokens low.

Compare with similar models

ModelDeveloperParametersContextLicenceInput / 1M
Mistral Small 24B (2501)Mistral AI24B32KApache-2.0$0.05
Cydonia 24B v4.1TheDrummer24B128K$0.42
Gemma 4 26B A4BGoogle DeepMind25.2B / 3.8B256KApache-2.0$0.07
gpt-oss-20bOpenAI21B / 3.6B128KApache-2.0$0.05

30 more models behind the same key and the same invoice.

Browse the catalog