Creating amazing AI models that are highly capable, yet tiny, with a focus on reasoning capabilities as well as coding, our models are highly adaptable.
Parallax Labs explores the creation of small models, often under 15 billion parameters, but with the capabilities of significantly larger models. Our models can reason like ones 100s of times their size. We are actively perfecting reasoning and coding capabilities with each model.
Our flagship reasoning model, with 9 billion parameters. Trained on over 150,000 quality reasoning, coding, and debugging traces. Experience top-tier Claude Opus reasoning locally, tool calling via MCP, and more. It is comparable to models like Claude Sonnet and GPT-5.6. What's the best part? Parallax 2 can run all locally.
The world's first local-first Agentic Development Environment. Built from the ground up to run Parallax 2 natively inside your IDE. No cloud APIs. No subscriptions. Just you, your hardware, and the most capable open-source coding agent ever built. Experience MCP tool calling, live web search, and native vision—completely offline.