Sakana AI
1 App · 10 Tools · 33 Intel
Sakana AI is a Tokyo-based frontier AI research and development company founded in 2023 by David Ha (CEO), Llion Jones (CTO), and Ren Ito (Chairman). Jones is a co-author of the landmark 2017 paper "Attention Is All You Need," which introduced the Transformer architecture, and Ha is a former research director at Google Brain. The company takes its name from the Japanese word for fish, reflecting its core philosophy of collective intelligence — developing nature-inspired AI approaches such as evolutionary algorithms and model merging rather than relying solely on brute-force scaling.
Is this you? Sign in with X to claim this profile.
Apps
Autonomous research agent that reasons for hours to produce structured strategy reports and options.
Tools
Agentic system that generates hypotheses, runs experiments, and writes scientific papers using progressive tree search.
Framework for automated scientific discovery in which LLMs propose hypotheses, run experiments, and write papers.
Hypernetwork approach that generates task-specific LoRA adapters for LLMs from a written task description.
Framework that lets language models adapt to unseen tasks in real time by composing expert modules on the fly.
Hypernetwork that converts a document into a LoRA, baking its factual knowledge into model weights.
Browser chat interface for the TinySwallow-1.5B model that runs entirely on-device with no API calls.
Multi-agent system behind one OpenAI-compatible API that orchestrates multiple LLMs on complex, multi-step tasks.
Python chat interface for the TinySwallow-1.5B model that runs entirely offline.
Japanese-tuned language model with a 256K-token context window, callable via Vercel AI Gateway.
Intel
Sakana AI's symposium report announces Sakana Sawara, a from-scratch agent-native LLM series; pre-training of the small model starts in December with weights planned for spring 2027. Also covers AI Scientist v3 and world-model robotics plans.
Sakana AI's TMLR paper scores LLM paper reviewers on catching planted contradictions, weighted by a claim-method-evidence knowledge graph. Its three-pass Multi-Layered Review agent outperforms the baselines tested, though all remain open to prompt injection.
Sakana AI's QD-Rec treats recommendation as open-ended search over item-hypothesis pairs. A contextual bandit picks items, an LLM mutates and crosses earlier hypotheses, and a quality-diversity archive keeps novel reasons. Beats ablations on LLM-judged quality; code released, human validation pendin
Sakana AI opens early access to its enterprise API Platform with OpenAI-compatible endpoints, workspace and role-based access control, audit logs and budget limits. Fugu and Fugu Ultra ship first, Namazu and Fugu Max follow. Pricing is shared only after application.
Sakana AI released Fugu Max, a cost-optimized orchestrator over open and specialized models priced at $2/$6 per million tokens, and Fugu Ultra v2, which claims top scores on several benchmarks without closed frontier models in its pool. Benchmarks are vendor-reported.
Sakana AI's PC-ALM generalizes predictive coding using an augmented Lagrangian instead of energy minimization, enabling local, backprop-free learning rules to train 1000-layer networks, a depth prior predictive-coding methods failed to scale to.
Sakana AI and UC Berkeley's MASS has one model act as executor, workflow optimizer and evaluator for multi-agent teams, then fine-tunes on the best team's runs. Two cycles on Qwen3.6-27B improved research benchmarks, with open code, and flags safety concerns about shared blind spots.
A Japanese research-meeting report on Sakana AI's ESCA system, which groups social posts into hierarchical narratives rather than keywords, flags AI generation and context tampering per post, and applies a game-search-style novelty search to surface small narratives before they spread.
Sakana AI details Fugu Max, which routes across an expanded pool of open and specialized models (including NVIDIA Nemotron) for frontier-level results at a fraction of the cost, and Fugu Ultra v2, which pushes peak orchestrated capability without leaning on frontier models directly.