My Favorites This Week
Gemini Native Image Generation
Mar 13, 2025
Gemini 2.0 Flash introduces native image generation, enabling contextual image creation, conversational edits, and long-text images—optimized for seamless chat interaction.
Google’s Gemma 3
Mar 12, 2025
Google’s Gemma 3 is Google’s latest set of lightweight, open AI models, built on Gemini 2.0 technology. Designed for speed and efficiency, they run directly on devices—from phones to workstations—empowering developers to build AI applications anywhere.
OpenAI’s Agents Building Tools
Mar 12, 2025
OpenAI introduced new APIs and tools to simplify agent development, including the Responses API, built-in web and file search, a new Agents SDK, and observability tools for workflow orchestration.
OpenAI New tools for building agents
Other Releases
Gemini 2.0 Flash Thinking Upgrades
Mar 13, 2025
Gemini 2.0 Flash Thinking boosts reasoning, expands app integration, and enables personalized AI. Key updates include free Deep Research, 1M token context, smarter task automation, and customizable AI experts (Gems).
Gemini Gemini Apps’ release updates & improvements
Reka Flash 3 Reasoning
Mar 11, 2025
Reka Flash 3 is a 21B general-purpose reasoning model, trained from scratch with synthetic and public data. Optimized for low latency and on-device deployment, it rivals proprietary models like OpenAI o1-mini.
Hunyuan-TurboS Model
Mar 10, 2025
Hunyuan-TurboS is the first ultra-large Hybrid-Transformer-Mamba MoE model, combining Mamba’s efficient long-sequence processing with Transformer’s strong contextual understanding. It overcomes traditional Transformer limitations in long-text training and inference.
👏 Thanks for reading!