My Favorites This Week
OpenAI’s 4o Image Generation
Mar 26, 2025
OpenAI’s new image generator for GPT-4o creates high-res images from text with advanced editing tools.
OpenAI Introducing 4o Image Generation
DeepSeek-V3-0324
Mar 25, 2025
Improved reasoning, better front-end development, and smarter tool-use capabilities.
Google Gemini 2.5 Pro Experimental Update
Mar 25, 2025
The new 2.5 models improve reasoning, accuracy, and coding skills, making them better at complex problems in science, math, and AI agents. 2.5 Pro delivers top-tier performance across benchmarks.
Google Koray Kavukcuoglu Gemini 2.5: Our most intelligent AI model
Other Releases
Manus AI Updates
Mar 28, 2025
Manus now has a mobile app, longer context, improved multimodal capabilities, and a more stable sandbox. All tasks run on Claude 3.7, with a premium subscription beta while keeping limited free access.
QVQ-Max
Mar 28, 2025
A powerful model that analyzes images and videos, solving math problems, coding, and even creating art.
Qwen Qwen Team QVQ-Max: Think with Evidence
Pika Flashback
Mar 28, 2025
Pika’s new feature lets you “Take a selfie with your younger self.”
Hyperbrowser AI’s HyperPilot
Mar 27, 2025
A playground for top AI browser agents like OpenAI’s CUA and Anthropic’s Claude Computer Use.
Y Combinator Launch YC: Hyperpilot: AI Browser Agent Playground by Hyperbrowser | Y Combinator
Qwen2.5-Omni-7B
Mar 27, 2025
A multimodal model that processes text, images, audio, and video while generating text and speech in real time.
Qwen Qwen Team Qwen2.5 Omni: See, Hear, Talk, Write, Do It All!
Microsoft Researcher
Mar 26, 2025
Combines OpenAI’s deep research model with Microsoft 365 Copilot for advanced multi-step research and deep search.
Ideogram 3.0
Mar 26, 2025
Enhanced photorealism, text rendering, and language understanding, plus new style reference features. Outperforms top image models.
Qwen2.5-VL-32B-Instruct
Mar 25, 2025
Improved human-like responses, better mathematical reasoning, and enhanced image understanding for detailed visual analysis.
Qwen Qwen Team Qwen2.5-VL-32B: Smarter and Lighter
Perplexity Answer Tab
Mar 25, 2025
New answer tabs let you search images, videos, travel, shopping, and more—all in one place.
Reve Image 1.0
Mar 24, 2025
Now the top-ranked image model, surpassing Midjourney and Imagen with stunning photorealism, precise prompts, and advanced text rendering.
👏 Thanks for reading!