NeurIPS 2025 - Google Code Golf Championship Based on ARC-AGI, create the shortest program that transforms input -> output $100,000 in prizes

Program synthesis ≠ training an LLM on piles of code. It’s something far more subtle, elegant, and powerful...

On Monday I had the pleasure of joining Forum on RTS - Radio Télévision Suisse to talk about Giotto.ai’s recent achievements! Watch the segment here : rts.ch/play/tv/forum/…

What do artificial intelligence and the atomic bomb have in common? Why is it crucial for Switzerland to attain AI sovereignty? Last week RTS came to our offices to find out. Watch the segment here : rts.ch/play/tv/19h30/… Read the transcript here: rts.ch/info/sciences-…

In a world obsessed with scale… 200M params score 24.58 on ARC-AGI-2. New high score 🚀🚀🚀

New ARC Prize 2025 High Score 24.58% by Giotto. ai (Aldo Podestà)

AGI progress bar: ███▒▒▒▒▒▒▒▒ 25%

ARC Prize Jeremy Berman Eric Pang How did they submit to v2? Is the score on the same data as the pubic lb on Kaggle? And now that grok has seen v2 problems, what prevents further progress using that info? I feel for people who achieve nearly the same score, within Kaggle compute limits. Their performance looks

Today: 25.83 on ARC-AGI. Tomorrow: a chance for Switzerland & Europe to lead in open, safe and smart AI.

Next Steps: - Get to 30% - Open source the solution - Generalise beyond ARC-AGI - Open source again

New ARC Prize 2025 High Score 27.08% by Giotto. ai (Aldo Podestà)

People still ask me how ARC-level reasoning translates to real intelligence. Simple: if a model gets 60%+ on ARC-AGI v2, it doesn’t get tricked by kindergarten word puzzles.

We are trying to go big of course... in case we fail we have an excellent plan b... pixel art

ARC Prize 2025: 30 Days Left Total Prizes Available: $1,000,000 ($125K Guaranteed this year) Just 30 days left (Nov 3) to make a submission (top score & papers) to ARC Prize 2025 on Kaggle

We’re at a decisive moment. The debate is no longer if Europe will build its own AI leaders, but who will take the lead. 👉 Full article (German, paywall): nzz.ch/wirtschaft/ein…

Can a small Swiss lab beat trillion-parameter AI giants? In our new Inside AI episode, Aldo Podestà of Giotto.ai shares how his team hit #1 on the ARC-AGI reasoning benchmark — with a model 1,000× smaller than Grok 4. 🎧Listen here: podcasts.apple.com/ch/podcast/con…

🎙️Great talking to Marcel Salathé on the EPFL AI Center podcast about Giotto.ai. Among other things, we talked about the origins of Giotto, what differentiates us from the other major players, and how R&D investments, even if risky, are absolutely crucial for real

Artificial Intelligence #AI / baby steps to AGI .. ARC - 2 Prize .. The race to reach “Artifical General Intelligence” just keeps accelerating. The next level is being defined now by the ARC -2 (Abstraction and Reasoning Corpus 2) competition and prize. 🏆

The LEADER on the leaderboard , Giotto.ai of Switzerland ( Giotto_ai) , led by Aldo Podestà has been delivering commercial LLMs with paying customers since 2018 - a full 4 years before ChatGPT was released. Full leaderboard here : kaggle.com/competitions/a…