ICYMI: Gemini 2.5 Flash Image Preview (nano-banana) is now available on AI Studio and APIs

Alibaba released Wan2.2-S2V - a 14B model designed for film-grade video making. Now available on the Wan website.

Claude Code in Github is now available in GA!

What’s new? 👀
- Trigger on more GitHub events (new issues, failed CI, custom conditions)
- Subagent support in actions
- Customizable templates for common workflows like code reviews

BREAKING 🚨: Looks like Nano banana is also rolling out to Gemini!

If I am not mistaken, image editing is now also available in the EU (it was restricted before)

BREAKING 🚨: Anthropic released "Claude for Chrome", a new browser agent that can take actions on your behalf as a research preview.
It will be released to the 1st 1000 users with Claude Max subscription.

Google introduces Gemini CLI integration in Zed editor

Google has integrated Gemini CLI into Zed, allowing developers to access AI-assisted coding, refactoring, and real-time support directly in the editor. The update supports complex tasks and aligns with open standards for extensibility.

🗞 #gemini@testingcatalog

BREAKING 🚨: xAI launched objects highlighting on Grok for iOS!

Object highlighting feature can point to certain objects in Grok Vision mode.

Grok got Astra before Gemini 🤯

Anthropic developing Claude Code web version to rival Codex

Anthropic is developing a web-based Claude Code with GitHub integration and managed sandboxes, targeting developers seeking browser tools for streamlined coding workflows. It positions Claude as a direct cloud IDE competitor.

🗞 #claude

Google Vids are about to get a bunch of new upgrades!
- A new image-to-video capability
- AI avatars! 👀
- Recording transcript cleanup

The basic Vids editor, without Gemini, is available to all consumers for free!

Moonshot AI released Kimi Slides! Now it is possible to generate a slide deck straight from the chat via Kimi website.

And it is real-time! Stunning to watch 👀

Now you can forward your emails to Manus AI so it can deal with them instead of you.

Cold emails are cooked 👀

OpenAI released a production-ready Realtime API for building voice agents and a new gpt-realtime model!

BREAKING 🚨: Microsoft introduced a new Copilot Audio Expression Labs project and 2 new models: MAI-Voice-1 and MAI-1-Preview.

- MAI-Voice-1 also powers Copilot Daily and Podcasts
- MAI-1-Preview is now available on LMArena and is soon to take over some chats

Krea AI is opening a waitlist for a new Real-time Video generation model.

"Behind the scenes, our approach is motivated by modern “world model” ideas: systems that learn how scenes evolve and how actions ripple forward in time."

Stitch by Google got a Canvas feature! Now you can plot and edit a whole UI journey over there.

So much to test 🤯

* Attaching a proper video

Looks like Microsoft is already rolling out the "Study and Learn" option for Copilot!

Stealth shipping 👀

OpenAI keeps experimenting with a "Thinking effort" selector for ChatGPT. In this experiment, "Juice 200", aka "Max thinking", is only available to Pro and Enterprise users.

Microsoft keeps working on its Library feature for Copilot by adding new categories for other artefact types like quizzes, podcasts and research results.