AI
Inception bets on diffusion language models with Mercury 3
Inception released Mercury 3, a diffusion-based language model it says runs at about 1,000 tokens per second and matches mid-tier offerings from OpenAI and Anthropic on quality. Launch pricing is $0.10 per million input tokens and $0.35 per million output, which the company describes as 50% below the eventual standard rate. Inception also credits swarms of the model itself with a 16% speed improvement during development.