Blog
FEATURED
Mistral releases Codestral Mamba for faster, longer code generation
The well-funded French AI startup Mistral, known for its powerful open source AI models, launched two new entries in its growing family of large language models (LLMs) today: a math-based model and a code generating model for programmers and developers based on the new architecture known as Mamba developed by other researchers late last year.
Mamba seeks to improve upon the efficiency of the transformer architecture used by most leading LLMs by simplifying its attention mechanisms. Mamba-based models, unlike more common transformer-based ones, could have faster inference times and longer context. Other companies and developers including AI21 have released new AI models based on it.
Now, using this new architecture, Mistral’s aptly named Codestral Mamba 7B offers a fast response time even with longer input texts. Codestral Mamba works well for code productivity use cases, especially for more local coding projects.
Mistral tested the model, which will be free to use on Mistral’s la Plateforme API, handling inputs of up to 256,000 tokens — double that of OpenAI’s GPT-4o.
In benchmarking tests, Mistral showed that Codestral Mamba did better than rival open source models CodeLlama 7B, CodeGemma-1.17B, and DeepSeek in HumanEval tests.
Developers can modify and deploy Codestral Mamba from its GitHub repository and through HuggingFace. It will be available with an open source Apache 2.0 license.
Mistral claimed the earlier version of Codestral outperformed other code generators like CodeLlama 70B and DeepSeek Coder 33B.
Code generation and coding assistants have become widely used applications for AI models, with platforms like GitHub’s Copilot, powered by OpenAI, Amazon’s CodeWhisperer, and Codenium gaining popularity.
Mathstral is suited for STEM use cases
Mistral’s second model launch is Mathstral 7B, an AI model designed specifically for math-related reasoning and scientific discovery. Mistral developed Mathstral with Project Numina.
Mathstral has a 32K context window and will be under an Apache 2.0 open source license. Mistral said the model outperformed every model designed for math reasoning. It can achieve “significantly better results” on benchmarks with more inference-time computations. Users can use it as is or fine-tune the model.
“Mathstral is another example of the excellent performance/speed tradeoffs achieved when building models for specific purposes – a development philosophy we actively promote in la Plateforme, particularly with its new fine-tuning capabilities,” Mistral said in a blog post.
Mathstral can be accessed through Mistral’s la Plataforme and HuggingFace.
Mistral, which tends to offer its models on an open-source system, has been steadily competing against other AI developers like OpenAI and Anthropic.
It recently raised $640 million in series B funding, bringing its valuation close to $6 billion. The company also received investments from tech giants like Microsoft and IBM.
Join our AI & tools
news weekly newsletter!
Latest Posts
News 1 (Apps)
TECH OpenAI sends internal memo releasing former employees from controversial exit agreements
OpenAI on Thursday backtracked on a controversial decision to, in effect, make former employees choose between signing a non-disparagement agreement that would never expire, or keeping their vested equity in the company.
Chris
May 23, 2024
News 2 (Products)
Google's AI Feature Suggested Using Glue to Keep Cheese on a Pizza
The tool, which gives AI-generated summaries of search results, appeared to instruct a user to put glue on pizza when they searched "cheese not sticking to pizza."
Chris
May 23, 2024
News 3 (Tutorial)
Meta Creates Group to Advise on AI Products
The Meta Advisory Group is composed of outside advisors that Meta's management team will periodically consult with on strategic opportunities related to our technology and product roadmap.
Chris
May 22, 2024
News 1 (Apps)
A new way to generate basketball analytics through tracking with computer vision and AI
In honor of the playoffs, I’d like to showcase what we’ve been working on here at Nexavision — a new way to generate basketball analytics through tracking with computer vision and AI:
Chris
May 23, 2024
News 2 (Products)
Amazon plans to give Alexa an AI overhaul — and a monthly subscription price
Amazon is planning to unveil a souped-up version of its decade-old voice assistant this year and will charge a monthly fee, sources say.
Chris
May 23, 2024
News 3 (Tutorial)
Adobe brings Firefly AI-powered Generative Remove to Lightroom
Adobe announced on Tuesday the addition of a Generative Remove feature for Lightroom. Built atop Firefly, the GenAI feature makes it possible to seamlessly edit objects out of photos. The feature arrives on Tuesday as early access.
Chris
May 22, 2024
:format(webp)/cdn.vox-cdn.com/uploads/chorus_asset/file/25379248/247075_Humane_AI_pin_AKrales_0120.jpg)
News 4 (Apps)
Humane is looking for a buyer after the AI Pin’s underwhelming debut
The startup apparently thinks it’s worth between $750 million and $1 billion despite the deep software flaws and hardware issues of its first product.
Chris
May 22, 2024
News 5 (Products)
Meta introduces Chameleon, a state-of-the-art multimodal model
The architecture of Chameleon can unlock new AI applications that require a deep understanding of both visual and textual information.
Chris
May 22, 2024
News 6 (Tutorial)
In Seoul summit, heads of states and companies commit to AI safety
Government officials and AI industry executives agreed on Tuesday to apply elementary safety measures in the fast-moving field and establish an international safety research network.
English 
Polish