Mistral introduces two new AI models: Ministral 3B and Ministral 8B. These models excel in knowledge, reasoning, and efficiency for on-device and edge use cases, supporting up to 128k context length. Ministral 8B features a sliding-window attention pattern for faster inference. They cater to local, privacy-first applications like on-device translation and autonomous robotics. Both models outperform peers in benchmarks and are available now.
Mistral AI announces Mistral Large 2, a new language model with enhanced capabilities in code generation, mathematics, reasoning, and multilingual support. It features a 128k context window, supports 80+ coding languages, and is designed for single-node inference. Released under the Mistral Research License, it excels in performance and cost-efficiency. Available on la Plateforme and through major cloud providers.
Mistral is introducing new AI model fine-tuning options, including paid plans, to let developers and enterprises fine-tune its generative models for particular use cases. Mistral has released an SDK, Mistral-Finetune, optimized for multi-GPU setups but scalable to a single GPU. It also launched fine-tuning services via API and custom training services for select customers.
Mistral, a French AI startup backed by Microsoft, has released Codestral, its first generative AI model for code. Codestral is trained on over 80 programming languages, including Python, Java, and C++. It helps developers write, complete, and debug code, and supports advanced AI applications. Codestral is available for download on HuggingFace and can be used via a dedicated API endpoint. It integrates with tools like VSCode and JetBrains for enhanced productivity.
Markifact is a no-code marketing automation platform that lets users create AI-driven workflows for automating marketing tasks. Users can trigger workflows, connect apps, and track performance through a visual interface. The platform offers pre-built templates, integrates with various marketing tools, and supports team collaboration.