Announcement
Character.ai is claimed to have doubled production inference throughput using AMD Instinct GPUs
AMD says Character.ai cut its cost per token by 50% using Instinct GPUs on DigitalOcean.
TLDR
AMD says Character.ai doubled production inference throughput and cut cost per token by 50% using Instinct GPUs on DigitalOcean. AMD also says Character.ai gained more room to serve users with the same compute footprint.
Combined views
29.3K
1 Source, first seen ago
