Inception releases Mercury 3 diffusion model, claims 8x speed

Inception AI released Mercury 3, a diffusion language model it says outperforms GPT-6 Luna and Claude Haiku 5.5 on eight of 12 internal benchmarks while running four to eight times faster. Inception measured 1,068 tokens per second on Nvidia GPUs, versus about 240 for Haiku 5.5 and 130 for GPT-6 Luna, with a 1 million token context window. Mercury 3 lists at 20 cents per million input tokens and 70 cents per million output tokens, with cached input 90% cheaper; a launch promotion halves both rates for now. Developers can use it through Inception's platform, OpenRouter, Baseten and Vercel, and OpenAI API code needs only minor changes. No outside group has published its own results yet.

AI-generated from reports · 2 hours agoUpdate