BREAKING 🚨: GEMINI 3.1 PRO HITS 77.1% ON ARC-AGI-2 AND 44.4%. ON HLE. ON PAR WITH OPUS 4.6 ON SWE BENCH VERIFIED.

ABSOLUTE SOTA 👀

Aerospace dashboards from Gemini 3.1 Pro!

Google releases Gemini 3.1 Pro with SOTA performance

Google has released Gemini 3.1 Pro, an upgraded AI model for science, engineering, and enterprise use. It delivers stronger reasoning with a 77.1% ARC-AGI-2 score, advanced code generation, 3D simulations, and complex system design, now available in preview across Google AI platforms.

🗞 #gemini

Claude Pro users can now use Claude in PowerPoint, as well.

Claude Pro is getting bigger every week 👀

Google has officially announced the Photoshoot feature in Pomelli.

"Available free of charge in the US, Canada, Australia & New Zealand!"

Gemini 3.1 Pro is now available on Perplexity 🔥

Google released CI Fixer for Jules SWE Agent, which automatically fixes CI issues to make it green.

Google test NotebookLM integration for Opal workflows

Google is testing NotebookLM integration within Opal, allowing notebooks to function as native workflow assets. Users could reference notebook data in Generate blocks, turning research into a persistent source for automated pipelines, alongside a forthcoming multi-model Agent feature.

🗞 #opal

OpenAI updated Code Blocks on ChatGPT to make them interactive. Users can now write, edit, and preview the code from the same inline widget.

Google will soon integrate NotebookLM into the Opal workflow builder. This also means that users will be able to rely on the context from their notebooks within their Super Gems on Gemini.

Microsoft develops Copilot Advisors to debate on any topic

Microsoft is developing Copilot Advisors, a feature that stages structured debates between two AI personas on user-defined topics. Users select domain-specific agents and positions, creating an audio-driven exchange designed to test ideas and support deeper analysis.

🗞 #microsoftcopilot

Anthropic released Claude Code Security in a limited research preview. Claude Code Security can automatically scan the codebase and suggest security fixes.

We have a new CSO now 👀

Claude Code on desktop can now start dev servers to preview the code and fix CI issues on their own.

This is what would let users drastically reduce back-and-forth follow-ups during vibe coding.

BREAKING 🚨: Pika Labs introduced Pika AI Selves, multi-faceted beings with persistent memory, who act as a living extension of you.

Available in early access 👀

Anthropic released Claude Code Security as research preview

Anthropic has launched Claude Code Security in limited preview, offering AI-driven code analysis to detect complex vulnerabilities and suggest patches for human review. Built on Claude Opus 4.6, it helps teams and open-source maintainers address critical security risks.

🗞 #claude

BREAKING 🚨: OpenAI is working on a new $100 Pro Lite subscription plan for ChatGPT, to capture a new segment of users between Plus and Pro.

h/t x@btibor91
* Not available yet

OpenAI prepares new ChatGPT Pro Lite tier priced at $100 monthly

OpenAI is reportedly preparing a $100 per month ChatGPT Pro Lite plan, positioned between Plus and Pro. The tier would target power users needing higher limits and potential agent-based features, while addressing the steep pricing gap in its current lineup.

🗞 #chatgpt