Skip to content
· NewsAvailable

Mistral AI launches Magistral-Small-2509: New 24B multimodal model

Mistral AI presents Magistral-Small-2509, a new 24-billion parameter multimodal AI model that integrates image comprehension and structured reasoning.

By the Aheadline editorial team·23 sep. 2026·2 min read·Source: Entity-watch: Mistral AIVerifierad signalAI-generated
Mistral AI launches Magistral-Small-2509: New 24B multimodal model
Mistral AI launches Magistral-Small-2509: New 24B multimodal model
By · Policy- & EU-reporter
Last updated
Vad betyder det för mig?

What happened?

Mistral AI has launched Magistral-Small-2509, a new multimodal reasoning AI model with 24 billion parameters. The model combines text and image comprehension through a new visual encoding module and is equipped with dedicated [THINK]/[/THINK] tags to structure its internal reasoning process.

Key facts

Parameterstorlek24 miljarder (24B)
AIME24-resultat86,14 % pass@1
GPQA Diamond-resultatÖver 70 %
Publiceringsdatum21 september 2026

Why it matters

By integrating image comprehension into a 24-billion parameter model, the barrier for complex multimodal applications is lowered. The model achieves 86.14% pass@1 on the AIME24 mathematics test—an increase of 15.6 percentage points compared to the previous version—and exceeds 70% on the GPQA Diamond knowledge test.

Who is affected?

The model is relevant to AI developers and companies looking to run advanced visual and text-based AI locally or on resource-constrained servers. The lower parameter count makes the technology more accessible to smaller players requiring high reasoning capabilities without extreme hardware requirements.

Impact on the EU

The model is developed by the French firm Mistral AI and is fully available to European and Swedish users. Since Mistral AI is headquartered within the EU, its development adheres to European regulations regarding data security and AI legislation (the EU AI Act).

What else you should know

Magistral-Small-2509 was released in a quantized 4-bit version via Unsloth (unsloth-bnb-4bit), making it possible to run the 24-billion parameter model on consumer hardware with limited VRAM.

Frequently asked questions

Quick answers about this story

Vad har hänt?
Mistral AI har lanserat den multimodala AI-modellen Magistral-Small-2509 med 24 miljarder parametrar och visuell bildförståelse.
När hände det?
Informationen publicerades den 21 september 2026.
Varför spelar det roll?
Modellen gör det möjligt att köra avancerad bild- och textanalys med hög resonemangsförmåga på mindre hårdvara, samt visar sina interna tankesteg.
Påverkar det EU?
Ja, eftersom Mistral AI är ett franskt bolag är modellen fullt tillgänglig i EU och utvecklad inom ramen för europeisk lagstiftning.
Original source
Entity-watch: Mistral AI·cf-1.com

The link opens in a new window and leads to the publisher's own site.

Verifierad signal

Källan har spårats automatiskt från utgivaren via Aheadlines signalkedja.

AI-verktyg i artikeln

Topics

#AI-benchmarking#Large Language Models (LLMs)#Mistral AI#Multimodal AI#AI-modell
[ STAY UP TO DATE ]

Get similar news straight to your inbox

No affiliate linksCancel anytimeGDPR-friendly
[ Frequency ]
[ What do you want to read about? ]

You'll receive updates on 2 topics.

The reader's room

Send in a question or an addition. The newsroom reads everything before it's published and replies when relevant. No AI-generated text – just people.

Sign in to submit a comment or question.

Loading comments…
How this affects you

Read the article through your role

  • Decide whether this affects strategy over 6–12 months or is just noise.
  • Discuss with leadership: do we own the right question or does ownership need to move?
  • Ask: what risk are we taking by NOT acting on this this quarter?

Generated angle — not editorial analysis of "Mistral AI launches Magistral-Small-2509: New 24B multimodal"