Sakana AI announced the release of four products: Sakana Chat (March 2026), Sakana Marlin, Sakana Fugu, and Sakana Translate, marking a shift from a research-focused company to also being a product…
Sakana AI, together with New York University, releases Dream-Cubed, a large-scale dataset of Minecraft worlds, and trains a family of transformers for generating interactive 3D environments at cube…
Sakana AI announces Fugu-Ultra v1.1, an upgrade to their orchestration engine that improves performance by up to 7.9 points over v1.0, beating leading models like Fable 5 in complex coding and…
Sakana AI announces its ICML 2026 paper 'UnMaskFork' which introduces a method for test-time scaling in masked diffusion language models by having multiple models collaborate through model switching…
Sakana AI released Fugu-Cyber, an update to its Fugu orchestration model, available as a new API endpoint.
Sakana AI is collaborating with NVIDIA to integrate NVIDIA's open model stack, including Nemotron, into Sakana Fugu, a multi-agent orchestration system, to advance open model innovation from Japan.
Sakana AI announces that a paper extending their collective intelligence research into physical hardware has been accepted for publication in Nature Communications.
Sakana AI, in collaboration with MIT and NYU, published a paper at GECCO 2026 titled 'In Search of the Ingredients of Open-Endedness: Replicating Picbreeder with Large Vision-Language Models'.
Sakana AI announced a research paper on a multi-agent coordination framework called Sheaf-ADMM, to be presented at ICML 2026.
Sakana AI announces a research paper at ICML 2026 that bridges parametric and nonparametric black-box optimization methods, showing they are variations of a single update equation.
Sakana AI announces the launch of Sakana Fugu, a new product that is a multi-agent orchestration system accessible via a single API. It comes in two models: Fugu and Fugu Ultra.
Sakana AI announced the formal establishment of the RSI Lab, a dedicated research group focused on recursive self-improvement technology using foundation models.
Sakana AI introduces DiffusionBlocks, a method to train neural networks one block at a time by reinterpreting the forward pass as a diffusion process, reducing memory requirements while matching…
Sakana AI, in collaboration with NVIDIA, announces a paper accepted at ICML 2026 introducing open-source GPU kernels and a new sparse packing format (TwELL) to speed up inference and training of…
Sakana AI introduces KAME, a tandem architecture for real-time speech-to-speech conversational AI that uses a fast speech model and an asynchronous backend LLM to enable 'speak while thinking'.
Sakana AI announces a research paper accepted at ICLR2026 on a Conductor model trained with reinforcement learning to orchestrate multiple AI agents via natural language, achieving state-of-the-art…
Sakana AI announces a new paper, 'TRINITY: An Evolved LLM Coordinator', accepted at ICLR2026. The paper introduces a system that coordinates multiple LLMs by dynamically assigning roles (Thinker,…
Sakana AI announces Sakana Fugu, a multi-agent orchestration system that coordinates frontier foundation models, now available for beta testing via API.
Sakana AI announces a prompting technique called String Seed of Thought (SSoT) that reduces output bias and increases diversity in LLM generations, accepted at ICLR2026.
Sakana AI releases Digital Ecosystems, a browser-based platform for interactive artificial life research where competing neural networks learn and adapt in a shared 2D grid.
Sakana AI announces that its paper on The AI Scientist, which describes an autonomous system capable of performing the entire machine learning research lifecycle, has been published in Nature after a…
Sakana AI announces Doc-to-LoRA and Text-to-LoRA, two research projects that use hypernetworks to generate LoRA adapters on the fly, enabling instant LLM customization without traditional fine-tuning.
Sakana AI announced a strategic partnership with Datadog to collaborate on joint research initiatives, open-source contributions, and go-to-market efforts, accelerating engagement with large…
Citi has made a strategic investment in Sakana AI, marking Citi's first such investment in a Japanese company, to accelerate innovation in global financial services.
Salesforce Ventures has invested in Sakana AI, highlighting the startup's research capabilities and understanding of Japanese enterprise needs.
Sakana AI announced the relocation of its head office to Azabudai Hills Mori JP Tower, citing business expansion and organizational strengthening.
Sakana AI announces a strategic partnership with Google, including a financial investment following the Series B round.
Sakana AI published an unofficial guide on what they look for when interviewing research candidates, emphasizing understanding over implementation, asking questions that distill the problem space,…
Sakana AI announces RePo, a method that allows language models to dynamically reorganize their context by learning to assign positions based on content relevance, improving robustness on noisy…
Sakana AI introduces DroPE, a method to extend the context length of pretrained LLMs by dropping positional embeddings during inference, requiring less than 1% of the original pretraining budget and…
Sakana AI, in collaboration with MIT, releases a research paper and technical report on Digital Red Queen (DRQ), an algorithm that uses large language models to drive adversarial evolution of…
Sakana AI's ALE-Agent achieved 1st place in the AtCoder Heuristic Contest 058, outperforming 804 human participants, marking the first known instance of an AI agent winning a major optimization…
Sakana AI announced it has raised 32 billion yen (200 million US dollars) in Series B funding to accelerate its mission of developing sustainable AI and implementing AI that benefits Japan.
Sakana AI introduces Petri Dish Neural Cellular Automata (PD-NCA), a new ALife substrate where multiple neural cellular automata learn to self-replicate and grow via ongoing gradient descent, leading…
Sakana AI announced the launch of its Business Development Division to begin providing services to customers, focusing on solving business and social challenges for Japanese companies using…
Sakana AI's ShinkaEvolve framework, an evolutionary code optimization tool powered by LLMs, was applied to optimize the SAT encoding in a solution for the ICFP Programming Contest.
Sakana AI and Daiwa Securities Group announced a long-term partnership to co-develop a 'Total Asset Consulting Platform' powered by Sakana AI's models, aiming to provide personalized financial advice…
Sakana AI introduces ShinkaEvolve, an open-source evolutionary code optimization framework that uses LLMs to discover new algorithms with state-of-the-art performance and unprecedented sample…
Sakana AI announces a new research paper, 'Competition and Attraction Improve Model Fusion', presented at GECCO'25, which proposes M2N2 (Model Merging of Natural Niches), a method that evolves…
Sakana AI announces the launch of its Applied Team (Business Development Division) in 2025, with a focus on social implementation of cutting-edge generative AI technology in the finance, defense and…
Sakana AI announces AB-MCTS, a new inference-time scaling algorithm that enables multiple frontier AI models to cooperate, achieving strong performance on the ARC-AGI-2 benchmark.
Sakana AI introduces Reinforcement-Learned Teachers (RLTs), a new method for training teacher models to generate explanations for student models by learning to teach rather than solve problems.
Sakana AI introduces ALE-Bench, a coding benchmark for hard optimization problems, and ALE-Agent, a coding agent that achieved 21st place among over 1,000 human participants in a live AtCoder…
Sakana AI introduces Text-to-LoRA, a hypernetwork that generates task-specific LoRA adapters from a text description, presented at ICML2025.
Sakana AI announces the Darwin Gödel Machine (DGM), a self-improving coding agent that rewrites its own code to improve performance on programming tasks, leveraging foundation models and open-ended…
Sakana AI has signed a comprehensive multi-year partnership agreement with MUFG Bank to develop AI systems for banking.
Sakana AI announced the Continuous Thought Machine (CTM), a new neural network model that uses neuron synchronization for reasoning, inspired by biological brains.
Sakana AI won the Innovative Spirit Award at the US-Japan Global Innovation Challenge 2025, a defense innovation competition hosted by the US Department of Defense's DIU and Japan's ATLA.
Sakana AI releases a new reasoning benchmark based on Sudoku puzzles, including a technical report, leaderboard, and GitHub repository.
Sakana AI announces that a paper fully generated by The AI Scientist-v2 passed double-blind peer review at an ICLR 2025 workshop, receiving an average score of 6.33, and that they have open-sourced…