Mercury Coder is a new diffusion large language model (dLLM) from Inception Labs, optimized for code generation. Thanks to its diffusion approach, it offers up to 10× higher speed than current autoregressive models, achieving over 1,000 tokens per second on the NVIDIA H100. It outperforms models such as GPT-4o Mini and Claude 3.5 Haiku in output quality, making it an efficient tool for developers.
