· 4 min read
ai-security
155 Leaked Responses: Inside Cloudflare's GLM-5.3 KV-Cache Bug
Cloudflare's GLM-5.3 deployment advertised a 1.05M-token context window backed by a 681K-token cache. The gap, plus an SGLang bug, leaked 155 responses across tenants. What the incident report actually says, and what it means for anyone running multi-tenant inference.
Read more