Cross-request KV cache exposure in the GLM-5.3 Workers AI model, and our remediation with SGLang
Comments
Ranked by a public rubric, not by votes: what a comment adds counts, not who wrote it. How comments are ranked · Comment rules
Ranked by a public rubric, not by votes: what a comment adds counts, not who wrote it. How comments are ranked · Comment rules
Summary
Cloudflare disclosed that between August 27 and September 9, 2026, 155 responses from its GLM-5.3 Workers AI model leaked data from other customers' requests, affecting only prompts over 681,280 tokens.