No single model will win every workload.
Voice, coding sub-agents, and search pipelines each require a different balance of intelligence, latency, and cost. That’s where diffusion LLMs shine.
Excited to build towards this future alongside @baseten.
To celebrate, we’re giving builders:
👉100M free Mercury tokens
👉10x higher rate limits
👉A faster, more capable Mercury 2
Start building → https://t.co/EnT7LcVcYS
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.
subagent — A separate agent spawned by another to do one scoped piece of work in its own context, returning only the result.
Why it matters
A new serving path plus a real free tier: 100M tokenThe chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.Full definition → and 10x rate limits make it cheap to test whether a fast diffusion model can carry your high-volume, low-complexity calls.
