Inception Labs launches Mercury Voice, a diffusion LLM with 2x lower latency than GPT-6 Luna and Claude Haiku 4.5

StefanoErmon · x · 2026-09-30

Inception Labs has released Mercury Voice, a diffusion LLM specialized for agentic voice applications. The company claims it delivers 2x+ lower latency than models including GPT-6 Luna, Gemma 4 31B, and Claude Haiku 4.5, while beating them across a range of voice benchmarks. Co-founder Stefano Ermon announced the launch, and enterprise access is via the sales team.

Original post →

More from Models

Models channel →