TechPrufer
Sign inGet started

Blogs›Mistral AI›Multimodal Safety Classification

Multimodal Safety Classification

Multimodal Safety Classification

5
posts
2024–2026

Shieldstral introduces a 3B open-weights multimodal safety classifier that outperforms models up to 7x its size by framing content moderation as a policy-adaptive question-answering task. It accepts plain-language policies at inference time, unifying text and image safety evaluation without retraining, and delivers calibrated safety scores efficiently. Le Chat now integrates this capability, allowing for document and image analysis powered by the new Pixtral Large multimodal model, enabling summarization. Pixtral 12B is a natively multimodal model trained with interleaved image and text data, excelling in multimodal tasks and instruction following while maintaining state-of-the-art text-only performance. It features a new 400M parameter vision encoder and a 12B parameter multimodal decoder based on Mistral Nemo, supporting variable image sizes and multiple images within a 128k token context window. Pixtral 12B demonstrates strong performance on multimodal reasoning benchmarks like MMMU and excels in chart understanding, document question answering, and image-to-code generation.

2026

Introducing Shieldstral. | Mistral AI

8/4/2026

Introduced Shieldstral, a 3B multimodal safety classifier. Developed a novel approach framing content moderation as a policy-adaptive question-answering task, allowing for plain-language policies at inference time. Unified text and image safety evaluation without retraining. Achieved strong performance on text safety, refusal detection, policy adaptability, and multimodal benchmarks. Built using Forge, Mistral AI's platform for training, aligning, and evaluating custom models.

2024

Mistral has entered the chat | Mistral AI

11/18/2024

This post details the integration of the Pixtral Large multimodal model into le Chat, enhancing its capabilities for document and image understanding. It highlights the model's ability to process large, complex PDFs and images, extract information, summarize content, and perform semantic understanding. The post also mentions the use of an experimental model in conjunction with Pixtral Large for these advanced features.

[Deprecated] Pixtral Large | Mistral AI

11/18/2024

Introduces Pixtral Large, a 124B multimodal model built on Mistral Large 2, offering frontier-level image understanding. It achieves state-of-the-art performance on MathVista, DocVQA, and VQAv2, and is competitive on MM-MT-Bench and the LMSys Vision Leaderboard. Pixtral Large features a 128K context window and demonstrates multilingual OCR and reasoning capabilities. Also announces an update to Mistral Large 24.11 with improvements in long context understanding, system prompt, and function calling.

Mistral Moderation API | Mistral AI

11/7/2024

Mistral AI releases a new content moderation API, which is the same API powering moderation in Le Chat. This API is an LLM classifier trained to classify text inputs into 9 categories and offers two endpoints: one for raw text and one for conversational content. The model is natively multilingual and provides performance benchmarks including AUC PR across policies.

[Deprecated] Pixtral 12B | Mistral AI

9/17/2024

This post announces Pixtral 12B, a natively multimodal model. It details its architecture, including a new 400M parameter vision encoder and a 12B parameter multimodal decoder based on Mistral Nemo. The model supports variable image sizes and aspect ratios, and can process multiple images within a 128k token context window. It demonstrates strong performance on multimodal reasoning benchmarks (MMMU), chart understanding, document question answering, and image-to-code generation, while also maintaining text-only benchmark performance. The post also notes that Pixtral 12B is deprecated and replaced by newer models.