DeepMind’s AlphaEvolve,GPT-4.1 launched in ChatGPT &Anthropic’s Claude's major upgrade

Sign Up here | Advertise with Us

Welcome Back, AI Geeks!

It’s time to stay updated with What’s Up in AI.

DeepMind’s AlphaEvolve is pretty amazing, it solves tough problems and even gets better at solving them as it goes. GPT-4.1 just launched in ChatGPT, making things faster and way smarter. And Anthropic’s Claude is about to get a serious upgrade, with the ability to use tools and fix its own mistakes on the fly.

In today’s issue:

🤯 Claude’s major upgrade incoming
🧠 DeepMind’s AlphaEvolve
📈 Just launched: GPT-4.1 in ChatGPT
⚡ 5-4-3-2-1 hits you just can't miss
😂 Your AI-Powered Laugh

Read time: 5 mins

LATEST IN AI TODAY

🧠 DeepMind’s AlphaEvolve: AI That Solves and Improves

• DeepMind unveiled AlphaEvolve, a coding agent built to tackle hard technical problems like chip design, data center efficiency, and even improving Gemini’s own training.

• When tested on 50+ open math problems, it solved 75% — and improved the best-known solution in 20% of cases.

• The model blends Gemini’s creativity with a scoring system that checks and refines outputs before finalizing them.

• It’s aimed at boosting AI development across Google, pushing toward smarter, self-optimizing systems.

• Still early, but AlphaEvolve could become a key tool in how AI builds and fixes itself.

🤖 GPT-4.1 Is Now in ChatGPT: More Power, Less Lag

• OpenAI just rolled out GPT-4.1 to Plus, Pro, and Team users — with free users getting access to 4.1 mini.

• GPT-4o is being retired for now, but 4.1 combines o1’s coding skills, o3’s reasoning, and lightning-fast performance.

• It’s smarter, faster, and better at following instructions, especially when it comes to tool use and code.

• First announced in April for developers, it’s finally landing in ChatGPT itself.

• The mini version brings big upgrades too — even free-tier users get a noticeable boost.

🤖 Claude Gets an Upgrade: Tool Use + Self-Checks Coming Soon

• Anthropic is prepping new Sonnet and Opus models that can reason, use tools, and even self-correct mid-task.

• The models will propose ideas, check them using external tools or databases, and revise before replying.

• Same goes for coding — they’ll run tests while building, fix bugs, and ship stronger final code.

• It’s a step closer to agentic AI: models that think, act, and verify.

• No launch date yet, but the drop is expected in the next few weeks — and it’s set to be a big one.

BRAIN BOOST 🧠 

QUICK HITS 💥 

It’s time for some quick hits that you’ll love. Let’s go in…..4-3-2-1 👇️ 

5- PRODUCTIVITY TOOLS

🌐 Blaze: Create stunning websites and prototypes in seconds without writing code

🎥 VidMe: Generate high-quality UGC-style videos instantly with AI

🚀 Co-Launchly: Get a personalized roadmap and strategy for your next big product launch

🧠 Hedy AI: Unlock real-time insights from your meetings and conversations

🎨 Playground AI: Design art, posts, logos, and more with an easy-to-use AI image creator

4-AI JOB OPPORTUNITIES

1. Writer - AI Researcher

2. OpenAI - Software Engineer, Inference

3. Siena - Senior Fullstack Engineer

4. IgniteTech - Artificial Intelligence Engineer

Want us to search for tech related jobs in your country/city ? Post a request 🔍

3-THINGS IN TREND

1. Bri Guy AI shows how he uses Midjourney to make photos that feel like old disposable camera shots.

2. Someone told Gemini to think only in emoji and it actually worked.

3. Reddit users are making wild alternate reality games with ChatGPT, and it just rolls with whatever story you feed it.

2-USEFUL LINKS

1. The AI Automation Blueprint: From Barista to Billionaire with Andrew Wilkinson from Tiny.com

2. Anysphere’s Eric Zakariasson just shared a guide for which models are best for which tasks in Cursor.

1-TOP ROBO NEWS

UC Berkeley Engineers Built a Humanoid Robot You Can Make for Under $5K 🤖🛠️

Engineers at UC Berkeley just dropped an open-source humanoid robot that you can build for under $5,000, a massive shift in a world where similar bots usually cost over $100K. It’s 3D-printable, can walk, hop, and be remotely controlled. Perfect for researchers, students, and anyone curious about robotics, the code and training tools are all freely available.

WHAT ELSE HAPPENED IN AI 📰

🧪 Meta’s New Model for Drug Discovery
Meta’s FAIR team just released an open-source model and dataset built on 100M calculations and 6B compute hours to accelerate drug and materials development.

🔊 Stability & Arm’s On-Device Audio AI
Stability AI and Arm unveiled a compact, open-source audio model that can generate 11-second clips — like drum loops or effects — directly on your smartphone.

🎬  Alibaba’s Text-to-Video Model Drops
Alibaba launched Wan2.1-VACE, a new model that turns images, clips, and text into videos — with precise editing available via text instructions.

🚌  Uber Launches Route Share
Uber’s “Route Share” shuttles will hit fixed stops every 20 minutes for cheaper rides — basically reinventing the bus with surge pricing.

🛡️ OpenAI Launches Safety Evaluation Hub
OpenAI is releasing a Safety Hub to track how well its models handle hallucinations, harmful content, and more — with regular public updates to follow.

Funding 💰️ 

• Legal AI giant Harvey could soon land a $250M funding round that would value the company at $5B, up from $3B just a few months ago. 

GET AN AI-POWERED LAUGH 😁 

Acquire new customers and drive revenue by partnering with us

What’s Up in AI is one of the most engaged & loved newsletter for professionals, AI enthusiasts & business owners with 35,000+ working at the world’s leading startups and enterprises. Companies like Hubspot, Artisan, Guidde & Gamma etc. feature their products in What’s Up in AI. You can learn more about partnering with us here.

And that’s a wrap!

Do drop a comment to tell us👇️ 

We’ll be back soon with more spicy takes on What’s Happening in AI, so stay tuned & share our newsletter with a friend using this link!

Stay Tuned! 🤖 
Team What’s Up in AI