StudyArena

Mercury 2

ResearchClosed weightsText

Mercury family model by Inception.

About the Mercury family

Inception Labs' commercial diffusion LLMs — Mercury 2 generates text via parallel coarse-to-fine refinement at 1000+ tokens/sec, mainly aimed at coding and latency-critical apps.

All Mercury models

Who makes it

InceptionUnited StatesPalo Alto lab behind Mercury, the first commercial diffusion LLMs.

← Back to all models