Microsoft Designer App, OpenAI GPT-4o Mini, New AI Architectures, and Mistral's Codestral Mamba

Microsoft Designer App, OpenAI GPT-4o Mini, New AI Architectures, and Mistral's Codestral Mamba

Welcome to our weekly newsletter ??, your go-to source for the latest developments and trends in Generative AI.

Each edition brings you a curated selection of impactful news, insightful analyses, and exciting advancements from the dynamic world of generative AI. Stay tuned for a concise and informative exploration of this rapidly evolving field.

1. Microsoft Launches AI-Powered Designer App Globally

Microsoft's AI-driven Designer app is now available worldwide on iOS and Android, exiting its preview phase. The Canva-like tool enables users to create various designs using text prompts, including stickers, cards, and collages.

Supporting over 80 languages, Designer offers free mobile and Windows versions. The app features customisable templates and AI-powered image editing capabilities. Microsoft plans to integrate Designer with Word and PowerPoint through Copilot, enhancing document creation workflows. Read more


2. OpenAI Launches High-Performance, Low-Cost GPT-4o Mini

OpenAI has introduced GPT-4o mini, a cost-efficient, high-performance model aimed at surpassing GPT-3.5 Turbo and rivals like Google’s Gemini 1.5 Flash and Anthropic’s Claude 3 Haiku. Priced at $0.15 per million input tokens and $0.60 per million output tokens, it offers substantial savings.

GPT-4o mini achieves an 82.0% score on the Massive Multitask Language Understanding (MMLU) benchmark, surpassing Google’s Gemini 1.5 Flash (77.9%) and Anthropic’s Claude 3 Haiku (73.8%)

The model scores 82% on the MMLU benchmark and includes advanced safety features. Set to replace GPT-3.5 Turbo for ChatGPT Plus and Teams subscribers, GPT-4o mini is poised for diverse applications, from customer support to creative writing. Read more


3. New AI Architecture Challenges Transformer Dominance

Researchers at Stanford, UC San Diego, UC Berkeley, and Meta have unveiled Test-Time Training (TTT) models, a potential successor to transformer-based AI. TTT models promise greater efficiency in processing vast amounts of data while consuming less computational power. This breakthrough could address the increasing energy demands of current AI systems. While still in the early stages, TTT models join other emerging architectures like state space models in the race to overcome transformers' limitations, potentially reshaping the future of generative AI. Read more


4. Mistral AI Unveils Codestral Mamba for Advanced Code Generation

Mistral AI has introduced Codestral Mamba, a new language model optimized for code generation. This model diverges from traditional Transformer models, offering linear time inference and the ability to handle sequences of infinite length.

Codestral Mamba, featuring 7.3 billion parameters, is available under an Apache 2.0 license and can be deployed using Mistral's SDK or TensorRT-LLM. It is tailored for code productivity, ensuring high performance in complex coding tasks. Read more


? Katonic Highlights


?? Introducing Katonic Companion: Your GenAI powered productivity booster

Katonic Companion is an enterprise-grade Generative AI solution that enables organisations to give employees managed access to some of the most popular Generative AI models while maintaining control over usage, costs, and data security.

Watch the video below to explore Katonic Companion's 5 key features that can help boost productivity!


Subscribe for more exciting AI updates in the future. Have a great weekend! ?



要查看或添加评论,请登录

社区洞察

其他会员也浏览了