Skip to content

Comment on Show HN: GPU-Accelerated Inference Hostingparent

Comments

The pricing appears to be static per model with a ceiling on the monthly request count, not charged per request.

Edit: Actually, I didn't spot the free tier of 1000 requests. I wonder how you avoid the problem of a lot of users leaving defunct/disused models running while still keeping them hot - presumably some kind of limit to the model count?

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.