Mistral

News from 7 October, local time, 6 October, the French company Mistral AI officially launched the public preview of the Mistral Large 4 model, short ML4, code name le Chonk. As of this date, the user will be able to experience the preview API on the Mistral Studio, and the model weight will be officially available for download at the end of this month. Officially, the top network security agencies, trusted partners and government regulators are now working together in a real business environment..

On September 8th, the French artificial intelligence start-up, Mistral AI, announced today that it had completed the D round of financing and raised €3 billion (about MOP 23,439 billion) from post-investment valuations of over €21 billion. The recipients of this financing are Samsung Electronics, Scaleup Europe Fund, PSG Equity. The scale of this financing is a historical record of European technology companies. This part..

July 16, 2011 - Mistral AI announced its Voxtral speech understanding models yesterday, local time. The family of models, available as open source, provides leading-edge, low error rate performance at a lower price point to support real-world speech intelligence production applications. The Voxtral family of models, derived from Mistral Small 3.1, includes Voxtral Small, a 24B version for production applications, and Voxtral Mini, a 3B version for local/edge deployments, in addition to...

July 15 (Bloomberg) -- Apple is seriously considering acquiring French AI startup Mistral, according to Bloomberg yesterday. Mistral AI is a European AI startup that has raised a total of 1.1 billion euros in seven rounds of funding (note: the current exchange rate is about 9.214 billion yuan). Bloomberg reports the company is in talks with investors to raise up to $1 billion (roughly Rs. 7.174 billion at current exchange rates). Over the years, Mistral has launched a range of large and small language models, with its o...

June 11 (Bloomberg) -- French AI lab Mistral announced Tuesday that it is entering the field of inferential AI models. On June 10, Mistral officially launched its first family of inferential models, Magistral, which solves problems in a step-by-step process designed to improve consistency and reliability in disciplines such as math and physics. This series of models is designed to improve consistency and reliability in disciplines such as mathematics and physics through step-by-step problem solving, similar to other inference models such as OpenAI's o3 model and Google's Gemini 2.5 Pro. The Magistral series consists of two versions: Magist...

Technology media outlet WinBuzzer published a blog post yesterday (May 21) reporting that Paris, France-based Mistral AI has launched Devstral, an open-source AI model for advanced software engineering tasks. Citing the blog post, 1AI said that the model was developed by Mistral AI in collaboration with All Hands AI, with a parameter scale of 24 billion, and is currently in the "research preview" phase. It is in "research preview" and released under the Apache 2.0 open source license, which allows developers and enterprises to use it for commercial purposes without restrictions...

May 8, 2011 - Mistral launched Medium 3, a mid-weight version of its Mistral 3 series of models, on May 7, local time, touting both cutting-edge performance and significantly lower cost with a wide range of enterprise features. Mistral claims that Medium 3 meets or exceeds Anthropic's Claude Sonnet 3.7's 90% performance in every benchmark, and costs only 0.4 per million Token output ...

On March 10, Google announced in a blog post on March 7 the launch of Gemini Embedding, an AI-based text processing model that is now integrated into the Gemini API. The model came out on top in the Massive Text Embedding Benchmark (MTEB), outperforming competitors such as Mistral, Cohere, and Qwen to become the most powerful text embedding model available today, Cohere, and Qwen, making it the most powerful text embedding model available today. Gemini Embedding...

March 8 - The Wall Street Journal reported yesterday (March 7) that Mistral AI, a French company known as the "European version of OpenAI," plans to embrace an open-source strategy and release a new model that goes beyond DeepSeek. Speaking at Mobile World Congress 2025, Mistral AI CEO Arthur Mensch said the company will adopt an open-source strategy to develop more powerful AI technology at a low cost, which will not only help the company to iterate quickly in the competition, but also promote global AI ...

March 7 news, Mistral AI local time 6 announced the launch of Mistral OCR API, and claimed that the API has the best performance in the field of OCR optical character recognition. Mistral AI said the API has excellent multimodal document processing capabilities, from PDF, slides, mathematical expressions, LaTeX academic documents and other complex document elements to extract text content, the results will be clearly presented in the form of Markdown files. Mistral OCR models can...

Jan. 23 (Bloomberg) -- Arthur Mensch, founder and chief executive officer of French AI startup Mistral AI, said in an interview with Bloomberg Television on Tuesday that the company is not for sale and is planning to go through an IPO. Arthur Mensch also said in the interview that the company is opening an office in Singapore to strengthen its presence in the Asia-Pacific region, and is continuing to grow its business in Europe and the United States. continuing to grow its business in Europe and the United States. Like its larger U.S. rivals OpenAI and Anthropic, the company was founded in 202...

Recently, Mistral AI announced the release of version 25.01 of its Codestral programming model, with officials emphasizing that the release features major improvements in terms of processing context length and code completion efficiency. Specifically, Codestral 25.01 increases the model's supported context length by 256,000 tokens, which is claimed to be able to effectively deal with large-scale projects and complex code generation needs, and the new version of the model also supports more than 80 programming languages, covering Python, Java, JavaScript ...

Nov. 19 - Mistral AI announced yesterday, Nov. 18, a new multimodal AI model, Pixtral Large, with 124 billion parameters, based on Mistral Large 2, and designed primarily for processing text and images. Pixtral Large is now available under the Mistral Research License and Commercial License for research, education, and commercial use. Pixtral Large is a Mistral ...

In a Nov. 8 blog post, tech media outlet WinBuzzer reported that Mistral AI has launched a new content review API to meet the growing demand for a secure online environment. Citing a press release, the content vetting API is based on a fine-tuned Ministral 8B model that categorizes content into nine categories, including hate speech, violence, and personal data breaches. The content review API supports Chinese, Arabic, English, French, German, Italian, Japanese, Korean, Portuguese, Russian and Spanish...

French AI startup Mistral AI has launched a new free package that lets developers fine-tune and build test applications using the startup's AI models, the company announced in a blog post on Tuesday. The startup has also dramatically slashed the price for developers to access its AI models via API endpoints and added image processing capabilities to its free consumer AI chatbot, Le Chat. Mistral has launched a free package through its API services platform, la Plateforme, which allows developers to experiment without cost, review...

TechCrunch reported yesterday (September 11) that French AI startup Mistral has released Pixtral 12B, the company's first multimodal AI big speech model capable of processing images and text simultaneously. The Pixtral 12B model has 12 billion parameters, or about 24GB in size, and the parameters roughly correspond to the model's problem-solving ability, with models with more parameters generally performing better than those with fewer parameters. The Pixtral 12B model is based on the text model Nemo 12B...

AI competition is becoming increasingly fierce. After Meta launched the open source Llama 3.1 model yesterday, French artificial intelligence startup Mistral also joined the competition and launched the next-generation flagship model Mistral Large 2. Model Introduction The model has a total of 123 billion parameters and is more powerful than its predecessor in code generation, mathematics, and reasoning, and provides more powerful multi-language support and advanced function calling capabilities. Mistral Large 2 has a 128k context window and supports languages including Chinese in…

Recently, French AI startup Mistral released a new coding model - Codestral Mamba. This model is not only fast, but also can process longer codes, helping programmers and developers improve their work efficiency. Mistral has accumulated a lot of fame in the field of open source AI, and the Codestral Mamba launched this time is even more eye-catching. Codestral Mamba is based on a new architecture called "Mamba", which is more efficient than the traditional transformer architecture. Its design allows the model to process...

According to a report by the Financial Times early this morning Beijing time, Paris-based AI startup Mistral AI has now raised 600 million euros in funding, with a valuation of nearly 6 billion euros (about 46.8 billion yuan before). The company, which is backed by Microsoft and Nvidia, has just been established for a year and is seen as a "potential challenger" to OpenAI. Since December (last year), the investment in the company has tripled its valuation. Mistral CEO Arthur Mensch said in an interview: "From the beginning...

Mistral, a French AI startup backed by Microsoft and valued at $6 billion, recently released its first generative AI model for coding, Codestral. The model is designed to help developers write and interact with code, supporting more than 80 programming languages including Python, Java, C++ and JavaScript. Mistral said in its blog post that Codestral can complete coding functions, write tests, fill in parts of the code, and answer questions about the code base in English. Although Mis…

According to a report by the Wall Street Journal today, French startup Mistral AI is about to reach an agreement to raise funds at a valuation of $6 billion (currently about 43.38 billion yuan), which is almost three times the valuation six months ago. In the financing last December, Mistral was valued at $2.15 billion. According to people familiar with the matter, existing backers General Catalyst and Lightspeed Venture Partners are expected to be the new…

According to the latest news, OpenAI's European competitor Mistral is in talks to raise funds at a valuation of up to $5 billion. Mistral, a Paris-based open source artificial intelligence startup, has been in discussions with investors to raise hundreds of millions of dollars at a valuation of $5 billion, according to a person with direct knowledge of the matter. The company only recently began to generate revenue, having raised $415 million in December last year at a valuation of $2 billion. Source Note: Image generated by AI, image licensing service provider Midjourney Mi…

Microsoft's layout in the field of AI has aroused the attention of many media and followers. The official announcement of a multi-year cooperation plan with the French company Mistral AI has triggered many speculations. Microsoft published an article on its Azure blog saying that through cooperation with Mistral AI, its large language model can be used on Microsoft's Azure AI supercomputing infrastructure. In addition, Azure AI Studio customers can also access the large language model Mistral Large. According to the Financial Times, Microsoft also...

Recently, the highly anticipated French AI company Mistral announced the launch of its AI chatbot Le Chat. This chatbot is seen as a strong competitor to ChatGPT and is designed to become a "conversation portal" for users to interact with Mistral's various models (such as Mistral Large, Mistral Small, and Next). Currently, Le Chat has opened Beta testing, and users can experience it by registering as a Beta tester. In addition, Mi…