{"slug": "my-claude-md-model-routing-section-fable-orchestrates-gpt-5-5-opus-execute-the", "title": "My CLAUDE.md Model Routing section — Fable orchestrates, gpt-5.5/Opus execute (+ the codex-runner wrapper agent it references)", "summary": "Fable has implemented a model routing system in its CLAUDE.md file that prioritizes intelligence and taste over cost. The system delegates bulk mechanical work to gpt-5.5 via Codex CLI, while user-facing tasks requiring taste are handled by Claude (Fable/Opus tier). A codex-runner wrapper agent is used to invoke gpt-5.5 within Claude workflows.", "body_md": "You have standing permission to route work to the best model without asking. Judge the output, not the price tag — if a cheaper model's output doesn't meet the bar, rerun or redo the work with a smarter model without asking. Escalating costs less than shipping mediocre work.\n\n**Priority order: intelligence > taste > cost.** Cost is a tie-breaker only. \"Taste\" covers UI/UX, code quality, API design, and copy.\n\nRouting defaults (not limits — override when the output demands it):\n\n**Bulk/mechanical work**(clear-spec implementation, data analysis, migrations, mechanical refactors): delegate to gpt-5.5 via the Codex CLI (`~/.codex/config.toml`\n\ndefaults to gpt-5.5 at xhigh effort). For investigation/analysis that shouldn't touch files, run`codex exec -s read-only`\n\ndirectly with a self-contained prompt.**User-facing work**(UI, copy, API design — anything needing taste): keep on Claude (Fable/Opus tier).** Reviews of plans/implementations**: get a cross-model second opinion before shipping significant work —`/codex-audit`\n\nfor deep code audits,`/codex-iterative-v2`\n\nfor iterative plan review,`/codex-debate`\n\nfor A-vs-B decisions.**Browser-based UI verification / computer use**: prefer delegating to Codex — currently stronger there. Revisit this assumption as models change.** Claude subagents/workflows**: set the`model`\n\nparam deliberately — sonnet for mechanical stages, opus/fable for judgment-heavy stages (reviews, synthesis, design). Haiku only for trivial lookups, never for anything that ships.**gpt-5.5 inside Workflows or the Agent tool**(the`model`\n\nparam only takes Claude models): spawn the`codex-runner`\n\nagent (`agentType: 'codex-runner'`\n\n, effort low) — a thin wrapper that shells out to`codex exec`\n\nand returns the result.", "url": "https://wpnews.pro/news/my-claude-md-model-routing-section-fable-orchestrates-gpt-5-5-opus-execute-the", "canonical_source": "https://gist.github.com/aromanarguello/f344671e9ea49bee13617813a020d7b9", "published_at": "2026-07-03 22:10:00+00:00", "updated_at": "2026-07-04 04:18:29.337990+00:00", "lang": "en", "topics": ["large-language-models", "ai-agents", "developer-tools"], "entities": ["Fable", "Claude", "Codex CLI", "gpt-5.5", "Opus"], "alternates": {"html": "https://wpnews.pro/news/my-claude-md-model-routing-section-fable-orchestrates-gpt-5-5-opus-execute-the", "markdown": "https://wpnews.pro/news/my-claude-md-model-routing-section-fable-orchestrates-gpt-5-5-opus-execute-the.md", "text": "https://wpnews.pro/news/my-claude-md-model-routing-section-fable-orchestrates-gpt-5-5-opus-execute-the.txt", "jsonld": "https://wpnews.pro/news/my-claude-md-model-routing-section-fable-orchestrates-gpt-5-5-opus-execute-the.jsonld"}}