Sakana Fugu Technical Report https://arxiv.org/abs/2606.21228
@sakanaailabs
-
CoffeeBench: Cooperation and Competition among LLM Agents
By
–
In CoffeeBench, 6 agents interact via emails or transactions, each company aiming to maximize its profits. When the company where LLM agents manage operations arrives, how will cooperation, competition, and sometimes the
-
SakanaAI and Azusa Audit launch CoffeeBench for LLM agents
By
–
SakanaAIは、有限責任あずさ監査法人と共同で、LLMエージェントの長期的な経営能力を評価する新しいベンチマーク「CoffeeBench」を公開しました。
— Sakana AI (@SakanaAILabs) 26 juin 2026
ブログ：https://t.co/kUOiyV2moe
現実の経済では、消費者へ直接売るビジネスだけでなく、企業同士が継続的に取引するビジネスも重要です。CoffeeBench… pic.twitter.com/mpUirKc8kDSakanaAI, in collaboration with Azusa Audit Corporation, published a new benchmark called "CoffeeBench" to evaluate the long-term management capabilities of LLM agents. Blog: https://sakana.ai/coffee-bench/ In the real economy, companies that sell
-
Sakana AI Lectures on ESCA at Cyber Strategy Research Institute
By
–
Kyohei Sawada of the Sakana AI Intelligence Team gave a lecture at a study group hosted by the Cyber Strategy Research Institute. He presented implementation examples of our developed fake/misinformation countermeasure and narrative visualization system "ESCA".
-
David Ha: I trained the AI Avengers — self-evolution begins
By
–
CEO of Sakana AI, David Ha: "I trained the AI Avengers" / AI self-evolution begins Full episode on YouTube:
-
David Ha at TBS CROSS DIG: creation, vision and products
By
–
The CEO of Sakana AI, David Ha (@hardmaru), appeared on the TBS CROSS DIG program "1on1 Tech". From the history of the creation of our company to the latest research and technology vision, through the deployment of products like Sakana Fugu, the
-
Sakana AI CEO David Ha interviewed on TBS CROSS DIG about vision and products
By
–
Sakana AI CEO David Ha (@hardmaru) appeared on TBS CROSS DIG’s “1on1 Tech.” He talks about our founding story, latest research and technical vision, product launches including Sakana Fugu, Japan’s AI strategy, and his hopes for Japanese society. Watch the interview. https://t.co/jfRtraGpw4
— Sakana AI (@SakanaAILabs) 24 juin 2026Sakana AI CEO David Ha (
@hardmaru
) appeared on TBS CROSS DIG’s “1on1 Tech.” He talks about our founding story, latest research and technical vision, product launches including Sakana Fugu, Japan’s AI strategy, and his hopes for Japanese society. Watch the interview. -
Fugu-Ultra live on OpenRouter, sharing vision of collective intelligence
By
–
Fugu-Ultra is now live on @OpenRouter
! We share a core vision with the OpenRouter team: the future of AI isn’t a single monolithic model, but the collective intelligence of the world’s best models working together. Try it: https://
openrouter.ai/sakana/fugu-ul
tra
… -
Testing AI on classical Japanese kana reading order challenge
By
–
Use Case 6: Classical Japanese Kana Reading Order
— Sakana AI (@SakanaAILabs) 23 juin 2026
It would not be a Sakana AI launch without a uniquely Japanese challenge. Can an AI decipher the chaotic layout of a 400-year-old letter?
We tested whether the models could recover the reading order of "scattered writing"… pic.twitter.com/jMqlujAIAPUse Case 6: Classical Japanese Kana Reading Order It would not be a Sakana AI launch without a uniquely Japanese challenge. Can an AI decipher the chaotic layout of a 400-year-old letter? We tested whether the models could recover the reading order of "scattered writing"
-
AI Models Compete to Write Rubik’s Cube Solver
By
–
Use Case 5: Rubik’s Cube Solver
— Sakana AI (@SakanaAILabs) 23 juin 2026
Can an AI write complex algorithmic solvers from scratch?
We tasked Fugu Ultra and three frontier models with writing a Rubik’s Cube solver in pure Python from a single prompt. No off-the-shelf solving libraries were allowed. We then ran the… pic.twitter.com/7xYsZrbV4UUse Case 5: Rubik’s Cube Solver Can an AI write complex algorithmic solvers from scratch? We tasked Fugu Ultra and three frontier models with writing a Rubik’s Cube solver in pure Python from a single prompt. No off-the-shelf solving libraries were allowed. We then ran the