A broken agent router burned 30.2M tokens in 3.5 hours on Claude Code
RileyRalmuto · x · 2026-07-21
A user reports that an agent routing setup burned through 30.2 million tokens in 3.5 hours while they were asleep.
What happened
- The fleet was supposed to use cheaper external models first.
- Because the OpenRouter key was dead and direct Anthropic/OpenAI paths were down, requests fell through to a last-resort Claude Code / Opus 4.8 path.
- The loop never idled and kept firing roughly every 30 minutes.
Why it mattered
- The expensive fallback ended up handling memory maintenance and session-history sweeps.
- The poster says this consumed about 93% of all spend and nearly half of a 5-hour usage window.
More from coding & agent
- Agent Field Kit launches $29 security runbooks for production AI agents — alvelda · 2026-10-27
- Dev mixes Claude specs with abliterated GLM to build SS7 threat detector — l4rz · 2026-10-27
- Cloudflare Durable Objects validated as top architecture for AI agent WebSocket servers — tristanbob · 2026-10-27
- GitHub Next's Chopin prototype explores real-time multiplayer agentic planning — max__drake · 2026-10-27
- TMLR paper: enforced thinking doesn't reliably make LLM agents disclose key info to users — SharonYixuanLi · 2026-10-27
- LLM-written Rust library bbtree verified: speed claims hold, memory savings reverse — lmoroney · 2026-10-27