Introducing ShinkaEvolve, an open-source approach to sample-efficient LLM-driven program evolution 🧬 Paper: arxiv.org/abs/2509.19349 Blog: sakana.ai/shinka-evolve/ Code: github.com/SakanaAI/Shink… The AI Scientist, Darwin Goedel Machine, and AlphaEvolve have fundamentally shaped

Super fun to watch our very own AI Scientist-v2 knocking off one of the 2024 predictions for AI Scientists!

Petri Dish NCA: some recent fun work with ivy and Sebastian Risi, presented at ALife2025. Each colour is it's own species. The simulation is differentiable and learning is continuous. More fun videos and an explanation here: iyzg.github.io/petri-dish-nca/

Proud to share that my team, Unagi, won the ICFP Programming Contest 2025! 🏆 Pushing SAT solvers to their limits was a lot of fun. This is our 8th win. Grateful to my teammates: いもす するめ Yoichi Iwata tos（とす） chokudai(高橋 直大)@AtCoder icfpcontest2025.github.io/prizes.html youtube.com/live/6-qhBSKxU…

Excited to release our new work: Petri Dish Neural Cellular Automata pub.sakana.ai/pdnca We investigate how multi-agent NCAs can develop into artificial life 🦠 exhibiting complex, emergent behaviors like cyclic dynamics, territorial defense, and spontaneous cooperation!

Who will take the challenge!?

Here is an overview of ShinkaEvolve by Robert Lange Slides: docs.google.com/presentation/d…

Check out my recent talk 📺 about our work Sakana AI on ShinkaEvolve 🧬, its contributions to Team Unagi's ICFP win 🥇, and the future and challenges of automated discovery! From Vibe Coding to Vibe Research/Optimization. Viewing LLMs as artistic 🎨 mutation engines

Leveling up vibe coding into vibe optimization 🚀

Many years in the making. Stay tuned. November 20th, 2025.

Where do GPT-5.1 Thinking and Kimi K2 Thinking stand in algorithm engineering? 🤖🧠 Our ALE-Bench says: 🔹 GPT-5.1 Thinking ≈ GPT-5 Thinking 🔹 Kimi K2 Thinking ≈ DeepSeek V3.1 Terminus OpenAI is still clearly a step ahead in this area. Kimi K2 Thinking is one of the best

Many have already witnessed Gemini 3 Pro's astonishing performance, but, one more from us. It has surpassed the GPT-5 series to become the new leader in our coding benchmark ALE-Bench!!! Congratulations to the Gemini team!!! Google DeepMind More details→sakanaai.github.io/ALE-Bench-Lead…

Excited to announce our MIT Press book “Neuroevolution: Harnessing Creativity in AI Agent Design” by Sebastian Risi (Sebastian Risi), Yujin Tang (Yujin Tang), Risto Miikkulainen, and myself. We explore decades of work on evolving intelligent agents and shows how neuroevolution can

I’m beyond excited to announce our MIT Press book on Neuroevolution! An HTML version is now available for free on neuroevolutionbook.com, with a print edition coming out later in 2026. Real intelligence is not static; it evolves. For decades, the field of neuroevolution has

沖縄で開催中のICONIP 2025の公開フォーラムに、Sakana AI Research Scientistの山田祐太朗 Yutaro Yamadaが登壇します。その他、ALife Institute 岡瑞起先生 Mizuki Oka/岡瑞起の基調講演やパネル討論もあります。ぜひご覧ください。 11/24 (祝) 13:00- 現地＋online（無料・要登録） iconip2025.apnns.org/forum/

One of the most exciting new AI companies is Sakana AI in our opinion. Their CTO and their Co-founder Llion Jones Llion Jones is now saying it's time to move beyond transformers even though he was one of the 8 original inventors at Google. They are now investigating the

Sharing a new bite-sized paper: Predictive coding network == cellular sheaf We put linear predictive coding networks into the language of sheaf theory, and use cohomology/Hodge decompositions to understand when recurrent PCNs actually learn.

At #NeurIPS2025, our team at Sakana AI will be presenting two posters on agents and inference-time scaling for challenging coding tasks: AB-MCTS and ALE-Bench. Unfortunately, I won't be there in person, but please stop by the posters to chat with my teammates! 🐟🐟🐟 1.