Inception Labs

Inception Labs builds large language models powered by diffusion rather than the autoregressive generation most LLMs rely on. Instead of producing text one token at a time, its diffusion LLMs (dLLMs) generate many tokens in parallel, which makes them several times faster and less than half the cost of conventional models while holding frontier-level quality. The diffusion approach also allows finer control over outputs, such as adhering to specific schemas and semantic constraints. Its team is drawn from researchers and engineers who have worked at Stanford, Google DeepMind, Meta AI, Microsoft AI, and OpenAI, among others.

The lab's flagship Mercury family brings this architecture into production, built around fast, low-cost inference without trading away output quality. On Venice, Inception Labs is represented by Mercury 2, a 2026 text model and currently the lab's only entry in the catalog.