Mistral AI launches Magistral Small 1.2 – bringing multimodal AI to personal computers
Mistral AI has released Magistral Small 1.2, a new 24-billion parameter AI model. It introduces visual processing and advanced reasoning capabilities for consumer-grade hardware.
What happened?
Mistral AI has launched Magistral Small 1.2, a new AI model featuring 24 billion parameters. Based on the Mistral Small 3.2 architecture, the model has been optimised using supervised fine-tuning (SFT) and reinforcement learning (RL). A significant development is the integration of a visual encoder, enabling the model to process and reason about both text and images within a multimodal environment.
Key facts
| Modellstorlek | 24 miljarder parametrar (24B) |
|---|---|
| AIME24 pass@1 resultat | 86,14 % (+15,6 procentenheter) |
| Hårdvarukrav (4-bit) | NVIDIA RTX 4090 (24GB) eller Mac med 32GB RAM |
| Basarkitektur | Mistral Small 3.2 |
Why it matters
The model represents a major advancement for smaller, more efficient models capable of complex reasoning. In the AIME24 benchmark, it reached 86.14% in pass@1, marking a 15.6 percentage point increase compared to version 1.1. Due to 4-bit quantisation, it can run on a single NVIDIA RTX 4090 (24 GB VRAM) or an Apple Silicon MacBook with 32 GB of memory, significantly lowering the barrier to entry for local AI usage.
Who is affected?
The release is aimed primarily at developers, researchers, and enterprises seeking to build or deploy advanced AI applications without relying on cloud-based APIs. Individual consumers and smaller organisations with access to consumer-grade hardware also benefit, as they can now execute a powerful multimodal model locally.
Impact on the EU
As a French company, Mistral AI develops its models in accordance with EU regulations, including the AI Act and GDPR. The ability to run Magistral Small 1.2 locally provides European companies with a distinct advantage regarding data privacy and compliance with strict data protection legislation.
What else you should know
The launch highlights a clear trend in AI development, with focus shifting from exclusively massive data centre models to more efficient, specialised models that can operate locally. By offering openness and support for the open-source ecosystem, Mistral AI is challenging the market's proprietary giants.
Quick answers about this story
Vad har hänt?
När hände det?
Varför spelar det roll?
Vilken hårdvara krävs för att köra modellen lokalt?
The link opens in a new window and leads to the publisher's own site.
Källan har spårats automatiskt från utgivaren via Aheadlines signalkedja.
AI-verktyg i artikeln
Topics
Get similar news straight to your inbox
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.
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 1.2 – bringing multimoda"