Opus 5 is hereOpus 5 has fallback to Opus 4.8 due to safety filters. It's the same price as Opus 4.8

Beginning July 20, Fable 5 will be included in all Max and Team Premium plans, at 50% of limits. Pro and Team Standard users will continue to have access to Fable via usage credits, and will receive a one-time $100 credit.

Running Fable 5 for everything is a waste of tokens. Here’s how to have Fable 5 review your entire project before another model executes: Step 1. Install the new Superpowers skill. Claude Code: /plugin install superpowers@claude-plugins-official Codex CLI: /plugins → search "superpowers" → Install Plugin Cursor: /add-plugin superpowers Step 2. Run this prompt on your project: "review this whole project before I hand execution to a cheaper model. use the superpowers skills for structure. 1. brainstorming skill first. before you review anything, ask me what this project is for, what good looks like, and what I am optimizing for. do not assume. 2. read everything. the code, the docs, the config, the folder structure. map what exists and how the pieces connect. 3. review across four lenses: • what is overbuilt or redundant and can be cut • where it is fragile, unclear, or likely to break • what the goal needs that is missing • where the structure fights the goal 4. writing-plans skill next. turn the review into a plan. for each change write what, why, which files, the order, and how to verify it worked. rank by leverage, highest impact first. 5. anything you are unsure about becomes an open question for me at the top of the plan. output one plan file a cheaper model can execute step by step without you in the room." Step 3. Hand the plan file to Opus or GPT-5.5. Let it execute. Fable has a high token cost. But you only use it for the thinking.

If you use Claude Code with Fable 5, here's a skill pack. It's called Finding Unknowns Here's what it does and how to set it up: These 8 skills help you surface blind spots, gaps, and wrong assumptions before any real code gets written. The 8 skills: → blindspot-pass: Claude explores unfamiliar territory and reports the questions you didn't know to ask → brainstorm: surfaces scope you might be missing before you lock in a direction → interview: Claude interviews you about the task to pull out the context you forgot to include → references: hands Claude visual or written references when words aren't enough → plans: flags the decisions most likely to change first, before you've built around them → notes: keeps a running log of every place the build deviated from the plan so the next attempt starts smarter → pitches: writes a pitch doc that answers the objections reviewers were about to raise → merge-quiz: proves you actually understand what changed before you merge it How to install: Option 1. npx (fastest) TERMINAL COMMAND npx skills add Neeeophytee/finding-unknowns-skills Option 2. Claude Code plugin TERMINAL COMMAND /plugin marketplace add Neeeophytee/finding-unknowns-skills /plugin install finding-unknowns-skills You can also tell Claude to install the entire skill for you: http://github.com/Neeeophytee/finding-unknowns-skills

There are two ways to do it, here’s how: Option 1, the official plugin: /plugin marketplace add openai/codex-plugin-cc /plugin install codex@openai-codex /reload-plugins, then /codex:setup That unlocks /codex:review, /codex:adversarial-review, and /codex:rescue. Claude writes, GPT critiques, you ship. Option 2, the proxy. One alias makes Sol your main model: alias claudex=‘CLAUDE_CODE_SUBAGENT_MODEL=gpt-5.6-sol CLAUDE_CODE_ALWAYS_ENABLE_EFFORT=1 claude –model gpt-5.6-sol’ Bonus: Claude Code lets you set subagent model and effort yourself. Sol Ultra runs can burn several times the tokens of a base run, so right-sizing delegated work saves real money.