← All model prices

Mercury 3: price, context window and licence

Diffusion language model at ~1,068 tokens/sec; context up from 260K in Mercury 2.5.

Where these figures come from

Taken from the OpenRouter model registry, which lists what providers charge per token, and converted to a per-million rate. Checked 30 Oct 2026. Where OpenRouter publishes no cache-read rate, the cell above is blank rather than inferred.

More on Mercury 3

• AI 101 — What is tokens per second (tok/s)?
• Deep Dive — Clover 35B: a diffusion LLM at Qwen parity, 680 tok/s
• Mercury 3: a diffusion LLM at 1,068 tokens/sec, $0.10 per million

Compare every model's price →