DEVOURED
Introducing Mercury 2.5
Inception Labs claims its new Mercury 2.5 model matches GPT-5.6 Luna performance while offering massive context windows and aggressive pricing.
What: Mercury 2.5 is a new diffusion-based language model featuring a 260K-token context window and output speeds of 1,107 tokens per second, priced at $0.04 per million input tokens.
Decoder
- Diffusion language model: A class of generative model that iteratively refines noise into structured data, similar to techniques used in image generation, now applied to token sequences.
- Tokens per second: A measure of inference speed indicating how much text a model can generate in real time.
Original article
Mercury 2.5 is the largest diffusion language model ever trained. It performs comparably to cost-optimized frontier models like GPT-5.6 Luna (Low), Gemini 3.5 Flash-Lite, and Claude Haiku 4.5. The model outputs 1,107 tokens per second on widely available Nvidia GPUs and has a 260K-token context window. At launch, Mercury 2.5 is 80% off at $0.04 per million input and $0.15 per million output.
Explore all 55 articles in today's digest
Fresh dev news, AI-digested daily. Never miss a trend at devoured.fyi