Fast Decisions. Continuous Evolution.
Fast decision models · Recursive self-improvement (RSI) · Agents
English · 简体中文
Website · Playground · Documentation · News · Hugging Face · Leaderboard
Valen is an open-source organization working on fast decision models, recursive self-improvement (RSI) and agents: models that decide quickly, agents that act on those decisions, and methods that turn experience into improvement.
| Project | What it does |
|---|---|
| Valen-Decision | Multimodal System One decision model, 0.8B to 27B. Text, images and video in; probabilities over your candidates out. Try it · Models · Docs |
| SDKs & integrations | Python, JavaScript and C++ SDKs, an agent skill for coding agents, and workflow integrations for n8n, LangChain, Dify and Coze. |
| JevalKit | Evaluation toolkit and leaderboard for decision models. |
| Roller | Agent trajectory collection across models and Harnesses, in disposable Docker containers. |
| Awesome-RSI | Papers, code, benchmarks and models on recursive self-improvement. |
Valen is named after evolutionary biologist Leigh Van Valen, who proposed the Red Queen hypothesis: keep adapting while everything around you evolves. Our Chinese name 万澜 (Wàn Lán) means countless waves, the wave-shaped V in our logo.
Contributions are welcome: open an issue or a pull request in any repository.