We are leveraging diffusion technology to develop a new generation of LLMs. Our dLLMs are much faster and more efficient than traditional auto-regressive LLMs. And diffusion models are more accurate, controllable, and performant on multimodal tasks.
Blog – Inception 10x free tokens on Mercury 2 | Claim yours 10x free tokens on Mercury 2. Claim yours Blog Product Aug 11, 2026 Mercury 2 for Search: Fast enough to run a hundred times per query Read story Product Aug 11, 2026 Mercury 2 for Search: Fast enough to run a hundred times per query Read story Product Jul 29, 2026 More builders. More throughput. Better Mercury 2. Read story Product Jul 29, 2026 More builders. More throughput. Better Mercury 2. Read story Product Jul 14, 2026 Mercury 2: the first reasoning model fast enough to pick up the phone Read story Product Jul 14, 2026 Mercury 2: the first reasoning model fast enough to pick up the phone Read story