How to cut Fable 5.2 token costs by 90%: Arc routing plus a lean CLAUDE.md
EXM7777 · x · 2026-11-12
A practical recipe to cut Fable 5.2's heavy default token usage by 90%, built on Rayline's newly released open-source model router ARC, which claims 97.5% of Opus 5.5 quality at up to 1/10th the cost by picking the right model and thinking budget per agent step.
The setup:
- Wire Arc into Claude Code: install raydock, run raydock auth login, then raydock claude; run the main agent on sage and every subagent/background task on swift
- Add to global CLAUDE.md: always write in ASD-STE100 simplified technical English, reporting only what's needed for decisions
- Keep CLAUDE.md under 200 lines — longer files get partially ignored; rewrite from scratch every 2-3 months
- Run /context in a fresh session to see what fills the window before typing; remove the biggest item and re-check
- Prefer CLIs over MCPs: commands return only requested lines, while MCPs dump entire results into the context window
More from coding & agent
- United Airlines to Ban Personal AI Agents From Booking Flights for Customers — Polymarket · 2026-11-12
- Agentic Search explained: a 4-search loop costs about 5x one search — and queries leak to more companies — adnan_hashmi · 2026-11-12
- AWS says thousands of autonomous AI agents wrote every line of AWS Emulator — SumitGup · 2026-11-12
- freeCodeCamp launches free 2-hour course on agentic coding with Codex — adnan_hashmi · 2026-11-12
- A criminally underrated resource comparing tool-calling reliability across models and providers — ricklamers · 2026-11-12
- Organizing 55k data points is easy: enterprise data prep for agents is the hard part — BenBajarin · 2026-11-12