Mercury Coder is now in @zeddotdev for next edit predictions.
Most LLMs generate code one token at a time. Mercury uses diffusion, refining the whole output in parallel — so suggestions land fast and keep you in flow.
Get 1 free month with the API key in the link:
https://t.co/w0UO8UWWGO
token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.
Why it matters
Zed users can switch the edit prediction backend to a diffusion model built for low latency, with a free month to judge whether parallel generation actually improves how inline suggestions feel.
