by MistralReleased 20 June 2025
Mistral Small 3.2 24B Instruct is the update to Small 3.1 and takes text and pictures, answering in text. Compared with 3.1 it follows instructions more closely, repeats itself less, stops endless replies and uses tools and fills in exact forms more reliably. It reads about 256 pages.
You pay for what each request uses. No subscription, no minimum.
One message, the ordinary use
Every rate
| Meter | ↑ you send | ↓ comes back |
|---|---|---|
| Text tokens | 24,96 DA | 66,57 DA |
per 1M tokens
Four companies run this model. OpenDunes sends each request to one of them, picked by price and speed. If it fails, the next one takes it. DA per 1M tokens: ↑ what you send · ↓ what comes back.
| Provider | ↑ Sent | ↓ Back | Cache read | Latency | Throughput | Uptime |
|---|---|---|---|---|---|---|
| DeepInfraUnited States | 24,96 DA | 66,57 DA | none | … | … | … |
| Parasail | 29,96 DA | 99,85 DA | 16,64 DA | … | … | … |
| VeniceUnited States | 31,20 DA | 83,21 DA | none | … | … | … |
| MistralEurope | 33,28 DA | 99,85 DA | 3,33 DA | … | … | … |
You pay the price of the provider that served your request. Prices read 15:20 · speed over the last 30 minutes · uptime over 3 days.
Real requests, real replies. Try any of them yourself — one click, pre-filled.
What's the total on this invoice, and when is it due?
invoice-scan.jpg · photoThe total is 42 300,00 DA, due 5 September 2026. Worth a double-check: the handwritten day could read as 8 instead of 3.
Try in Chat LabTurn this customer email into JSON with intent, urgency and order_id.
customer-email.txtTry your own request — same model, pre-selected.
Open Chat LabSwap the base URL and the key. Your OpenAI code runs unchanged.
import os
from openai import OpenAI
client = OpenAI(
base_url="https://opendunes.com/v1",
api_key=os.environ["OPENDUNES_API_KEY"],
)
resp = client.chat.completions.create(
model="mistralai/mistral-small-3.2-24b-instruct",
messages=[{"role": "user", "content": "Two lines to welcome a shop in Oran."}],
)
print(resp.choices[0].message.content)Runs this request once on your balance.
Standard OpenAI fields, grouped by what they control. Anything not listed is accepted and ignored.
Shaping the reply
How long, how varied, where to stop.
max_tokenstemperaturetop_ptop_kmin_pstopseedfrequency_penaltypresence_penaltyrepetition_penaltylogit_biaslogprobstop_logprobsStructured output
Get JSON you can parse instead of prose.
response_formatstructured_outputsTools
Let the model call your functions.
toolstool_choiceAbout 0,058 DA per message. ↑ 24,96 DA · ↓ 66,57 DA per 1M tokens.
No. You pay per use from your balance, with no subscription. Try it in Chat at the same price.
Mistral Small 3.2 24B Instruct is the update to Small 3.1 and takes text and pictures, answering in text. Compared with 3.1 it follows instructions more closely, repeats itself less, stops endless replies and uses tools and fills in exact forms more reliably. It reads about 256 pages.
Mistral Small 4 costs 0,092 DA more for the same message.
Deposit dinars with your Edahabia or CIB card. Every model, Mistral Small 3.2 24B included, is paid from that balance. No foreign card, no conversion.
Mistral's model card tags 24 languages, Arabic and French among them. Darja is not named.
It reads text and images. Up to 256K tokens in one request, about 250 pages. It writes up to 12 000 words at once.
Yes, its weights are public. Open weights on Hugging Face
Mistral. Released 20 June 2025. It knows the world up to 31 October 2023.
Mistral's flagship speech-to-text model that turns audio into transcripts with word timestamps and speaker labels, with context biasing for domain-specific vocabulary
The Voxtral Mini derivative that only transcribes, taking audio from meetings, voice notes and podcasts and returning the text through a standard transcription request
Mistral's compact speech-to-text model that turns audio into transcripts with word timestamps and speaker labels, and can be steered with context biasing to catch domain terms
Mistral AI's text embedding model from late 2023, whose 1024-dimension vectors capture how passages relate, for semantic search and RAG; Codestral Embed is the version for code
From what its maker and two benchmark sites publish, and where it stands on OpenDunes. Read 7 Nov 2026.
Thinks first No.
Works through the problem before it answers.
Uses your tools
Can call functions your code provides.
Fills your JSON
Answers in the JSON shape you give it.
Searches the web No.
Can look things up online while it answers.
Cheaper on repeats No.
A prompt it saw a moment ago costs less the second time.
Open weights
The model's weights are public.
Knows events up to 31 Oct 2023
What it learned stops at this date.
Run a support assistant that looks up orders
Uses your tools
Read a product photo and fill a JSON listing
Images · Fills your JSON
Translate customer messages into French or Arabic