Get Zlio on your home screenOne tap, full screen, like an app.
Frontier Labs · first seen 20 October 2026, 15:41 UTC
Inception's Mercury 3 diffusion model claims 8x faster text
Diffusion-based text generation could shift LLM inference economics toward much lower latency.
Inception released Mercury 3, a diffusion-based text model it says generates output up to 8x faster than rivals and beats GPT-6 Luna and Claude Haiku 5.5 on some measures.
