Developers use a /goal command to prompt continuous execution.
Thariq#544
Claude Fable 5 changed how we work on the Claude Code team day to day.
We used to verify that Claude did the work right. Now we verify that it's doing the right work.
Here’s the 3 biggest changes:
Many users praise Claude Fable 5 for shifting Anthropic teams from task checks to goal alignment and enabling massive code migrations, while some dismiss the announcements as hype and criticize limited access.
Fable is a step-change in models, and I hope it changes how you work with Claude.
More to come in a series of posts on how it’s reshaped our work, but the TLDR: it’s time to be more ambitious.
Claude Fable 5 changed how we work on the Claude Code team day to day.
We used to verify that Claude did the work right. Now we verify that it's doing the right work.
Here’s the 3 biggest changes:
"We used to check if Claude is doing the work right, e.g. by double-checking its output, catching when it stopped early etc. With Claude Fable 5, I instead check if Claude is doing the right work"
- Thariq (@trq212) Claude Code
@ClaudeDevs I’m Putting it to the test.
@ClaudeDevs Why can't I see it in the desktop app yet?!
I have no patience lol.
This is one hell of a statment:
"We used to verify that Claude did the work right. Now we verify that it's doing the right work"
Some really cool recommendation for pushing Claude Code to its full potential. By Thariq (@trq212) from Claude Code team.
(Noted from his video by Grok)
- Shift from verifying whether Claude did the work right to verifying whether Claude is doing the right work.
- Treat Claude Fable 5 like a true thought partner by giving it the full context it needs upfront, rather than jumping straight into implementation.
- Involve Claude early in the thinking process by starting with a small spec and asking it to interview you about the implementation details before finalizing the spec file.
- Ask Claude to explore multiple directions for an idea and generate quick mockups (such as in HTML) for review, which helps catch misalignment before any code is written.
- Provide Claude with rich context instead of rigid constraints—for example, explain that a feature is an experiment likely to be deleted in a month so it avoids building anything painful to throw away.
- Give Claude explicit goals and verification methods once the direction is clear, especially for ambitious problems.
- Use the new /goal command in Claude Code, which helps the model keep working until the objective is fully complete.
- Use Workflows in Claude Code to let the model parallelize tasks, verify its own output, and prepare a report on what was implemented versus what differed from the plan.
- Prompt Claude with a combined instruction such as: “Set a goal to implement the spec fully, then use a workflow to verify each part of the plan, and prepare a report on what was implemented and if anything differed.”
- Be far more ambitious with Claude Fable 5 by assigning it tasks previously assumed to be impossible for LLMs, as the model now runs for hours, self-tests, and often produces higher-quality code than manual efforts.
Experiment boldly—for instance, I edited this entire video using Claude Fable 5—because the model raises the bar on what developers can realistically achieve in a single session.
"We used to check if Claude is doing the work right, e.g. by double-checking its output, catching when it stopped early etc. With Claude Fable 5, I instead check if Claude is doing the right work"
- Thariq (@trq212) Claude Code
If Fable is the safer public version, the real question is not only how strong Mythos is…As AI moves into biology and chemistry, how much of this path will remain human-led? AI can help cure biological “bugs” — but it may also discover vulnerabilities in human systems we don’t expect. AI + biology is the right frontier, but not with ungrounded silicon intelligence alone
@jxnlco yeah lol, is it obvious
I did write my own script though
@Layton_Gott @ClaudeDevs Just type /model claude-fable-5 and it will appear！
@Layton_Gott @ClaudeDevs I didnt see it eihter but use /model claude-fable-5 worked for me no problem.
@ClaudeDevs @cursor_ai bench data here is impressive for score but even more for the drop in cost per intelligence.
Fable5 Med $8.27/69.8% vs Opus 4.7 Max $11.02/64.8% Fable5 Low $5.70/64.2% vs Opus 4.8 Max $7.59/63.8%
The trend about the price per intelligence dropping is continuing.
@ClaudeDevs It's refusing to scan our apps for security vulnerabilities sadly
@ClaudeDevs Had it scan a project for bugs and it found a minor security thing. Instantly shut down... and now that session is toast because the security thing is in the context.
I read the docs as saying I'd be downgraded to Opus?
@ClaudeDevs Looks like I am not testing Fable much 😭
@ClaudeDevs This is the real shift imo. Once "did it do the task right" is solved, the hard part becomes deciding what the right work even is. Same thing slows enterprise rollouts — the bottleneck moves from the model to workflow + review design. Which of the 3 moved the needle most?
@ClaudeDevs You killed all my excitement for it immediately by telling me you ar gonna take it away in a few weeks
Honestly the disrespect towards loyal sub users is very disheartening, we end up getting the short stick every time
@ClaudeDevs Fable as planner and code reviewer is about all we can afford.
@ClaudeDevs Wish you came out with this sooner. Pretty burnt out by now with mediocrity, dishonesty, and laziness at scale with older models until now. Let’s hope this one is actually better this time instead of more marketing.
@ClaudeDevs
@ClaudeDevs Testing it now on the same real task I gave Opus earlier today.
Early impression: I’m correcting less and trusting it more. Let’s see if that holds through the final result.
Developers use a /goal command to prompt continuous execution.
Thariq#544Claude Fable 5 changed how we work on the Claude Code team day to day.
We used to verify that Claude did the work right. Now we verify that it's doing the right work.
Here’s the 3 biggest changes: