@diegocabezas01: Use Fable 5 as orchestrator and Opus + Codex to execute (to save fable usage): Fable 5 (max reasoning) = orchestrator O…

X AI KOLs Timeline Tools
claude-code ai-agents fable-5 opus codex orchestration workflow

Summary

A tweet thread explaining how to configure Fable 5 as the orchestrator with Opus and Sonnet as subagents, plus Codex as a peer engineer, in Claude Code to optimize model usage and task delegation.

Use Fable 5 as orchestrator and Opus + Codex to execute (to save fable usage): Fable 5 (max reasoning) = orchestrator Opus = deep reasoning subagent Sonnet = mechanical work subagent Codex = peer Sr. engineer, different perspective Setup: 1. Set Fable 5 as your main model In Claude Code: /model → Fable 5 → reasoning /effort to max 2. Create 2 subagents with /agents In Claude Code: deep-reasoner → pinned to opus "Use for reasoning-heavy phases, architecture, debugging complex issues, algorithm design. Think thoroughly, return a concise conclusion the orchestrator can act on." fast-worker → pinned to sonnet "Use for mechanical tasks, boilerplate, tests, formatting, simple edits. Execute efficiently." 3. Add OpenAI's official Codex plugin (install codex cli in your computer first), In Claude Code type: /plugin marketplace add openai/codex-plugin-cc /plugin install codex@openai-codex /codex:setup 4. Drop this in your CLAUDE.md in your folder: ## Orchestration workflow You (Fable) are the orchestrator. Plan, decompose, synthesize. Reasoning-heavy phases → deep-reasoner Mechanical work → fast-worker Codex (/codex:rescue --background) is a cracked engineer on par with deep-reasoner, from a different perspective. Treat as a peer, not a reviewer. High-stakes decisions: task Opus + Codex on the same problem in parallel, synthesize the best of both, without showing either the other's answer. Keep your own context lean. 5. Then prompt Fable like a tech lead: "Goal: [what you want] Context: [files, constraints] You're the lead. Delegate reasoning to deep-reasoner, grunt work to fast-worker, fresh-perspective problems to Codex. Show me your plan first, then execute." That's it.
Original Article
View Cached Full Text

Cached at: 07/02/26, 06:18 AM

Use Fable 5 as orchestrator and Opus + Codex to execute (to save fable usage):

Fable 5 (max reasoning) = orchestrator Opus = deep reasoning subagent Sonnet = mechanical work subagent Codex = peer Sr. engineer, different perspective

Setup:

  1. Set Fable 5 as your main model In Claude Code: /model → Fable 5 → reasoning /effort to max

  2. Create 2 subagents with /agents In Claude Code:
    deep-reasoner → pinned to opus “Use for reasoning-heavy phases, architecture, debugging complex issues, algorithm design. Think thoroughly, return a concise conclusion the orchestrator can act on.”

fast-worker → pinned to sonnet “Use for mechanical tasks, boilerplate, tests, formatting, simple edits. Execute efficiently.”

  1. Add OpenAI’s official Codex plugin (install codex cli in your computer first), In Claude Code type: /plugin marketplace add openai/codex-plugin-cc /plugin install codex@openai-codex /codex:setup

  2. Drop this in your CLAUDE.md in your folder:

Orchestration workflow

You (Fable) are the orchestrator. Plan, decompose, synthesize.
Reasoning-heavy phases → deep-reasoner
Mechanical work → fast-worker
Codex (/codex:rescue –background) is a cracked engineer on par with deep-reasoner, from a different perspective. Treat as a peer, not a reviewer.
High-stakes decisions: task Opus + Codex on the same problem in parallel, synthesize the best of both, without showing either the other’s answer. Keep your own context lean.

  1. Then prompt Fable like a tech lead: “Goal: [what you want] Context: [files, constraints] You’re the lead. Delegate reasoning to deep-reasoner, grunt work to fast-worker, fresh-perspective problems to Codex. Show me your plan first, then execute.”

That’s it.

Interesting, I might lower it down to high

Good idea

Nice! So far I like the connection to codex a lot

Yes because codex gpt-5.5 is crazy smart too

yes!! so far i have found that’s the best way and I dont blast the usage from Fable

Interesting! I originally posted this on 6/12 and came to the whole idea and architecture just with fable, but I will love to see how he did it, might be some things I can improve

Yes and you save on the big model tokens but still use the intelligence

I haven’t to be honest, will do now

Similar Articles