Smaller, faster, safer: running Kimi and GLM at scale

https://blog.cloudflare.com/smaller-faster-safer-models/