Inception "Mercury 3", 1000 tok/s diffusion LLM

Inception has released Mercury 3, a diffusion-based large language model that generates text at 1,000+ tokens per second while matching or exceeding the quality of faster frontier models like GPT-6 Luna and Claude Haiku 5.5. The model supports 1 million token context length and costs $0.10 per million input tokens (50% off launch price), demonstrating that diffusion LLMs can scale effectively for commercial applications. Early users are deploying Mercury 3 for real-time applications including browser agents, interactive tutoring, and customer support.