Claude Fable 5.2 Is Here, and Pro Gets a Free Week
Anthropic released Claude Fable 5.2, its most capable model, and Pro subscribers can use it inside their plan until November 16. Here's who gets what, how Fable resets work, and how to test it against Opus 5.5 on your own work this week.
AI for Mortals Staff
Anthropic released Claude Fable 5.2, its most capable model, in the Claude apps and Claude Code. It writes its answers up to 30% faster than Fable 5.1, and Anthropic's testers found it explains its work more clearly than earlier Fable models.
For one week, Claude Fable is also open to Pro subscribers inside their regular plan. That window closes on November 16, so if you're on Pro, this is the week to try it.
What Claude Fable 5.2 can do, and what changed since 5.1
Fable is the top of Claude's lineup, above Opus. The numbers run on separate tracks, which is why Fable 5.2 sits above Opus 5.5.
In Anthropic's launch video, Claude takes a request to find out why sales fell for a quarter, finish a brief and build a small app to view the spreadsheets. It splits the job between two helpers, extra copies of Claude that each take a part of the job at the same time, then tells you in plain sentences which helper is doing what.
Anthropic's examples of what it did in testing are big ones. In one experiment, an Anthropic engineer had Fable 5.2 run more than 1,000 helpers for eight days, and it produced a working operating system that boots on a laptop. Another prompt had it design three electronic devices, circuit boards included, that all worked without any electrical fixes. It also controlled robot arms to solve a puzzle cube, open a bottle and draw a picture.
Claude Mythos 5.2 is the same model without the safety filters that refuse hacking and biology-lab requests. Only security and medical-research groups that Anthropic approves through its verification programs can use it.
Fable keeps those filters, and Anthropic says they now block about 55% fewer requests than Fable 5.1 did. When one does get blocked, an older model answers in its place: Opus 4.8 for cybersecurity questions, Opus 5 for biology.
Who can use Claude Fable 5.2, plan by plan
According to Anthropic's Help Center:
- Max and Ultra: Fable is part of your plan, for up to half of your weekly limit. You also get one to three Fable resets, explained below.
- Pro: Fable normally runs on usage credits you pay for separately, a setup that started when Fable 5 moved off Pro in July. Through November 16 it runs inside your plan instead, plus one reset.
- Team: premium seats get what Max gets, and standard seats get what Pro gets.
- Free: no Fable.
If you're not sure which plan fits you, our guide to what Claude's Free plan includes and what Pro, Max and Ultra add covers all four, and Ultra is the plan that replaced Claude's $200 tier.
How the free week of Claude Fable on Pro works
From launch until November 16 at 11:59 pm Pacific, Pro subscribers can use any Fable model, including 5.2, inside their regular weekly limit. You can spend up to half of that weekly limit on Fable, the same cap Max and Ultra have.
Once you've used that half, or once the week ends, Fable goes back to billing your usage credits, which you pay for on top of your plan, if you've turned them on. If you haven't, switch back to Opus. Anything you didn't use of the half doesn't carry over.
Your one Pro reset works on top of this: once you've spent your Fable half, a reset gives it back so you can use Fable again before November 16.
Pick it from the model menu in the Claude app or in Claude Code, the same way you'd switch to Opus.
How Fable resets work, and when to use yours
A Fable reset gives back the Fable usage you've spent this week. Anthropic's example: if you've used 80% of your weekly limit, 50% on Fable and 30% on other models, a reset takes you back to 30%. If your plan has a five-hour limit, it refills that completely too.
How many you get depends on how long you've paid for Claude:
- Max and Ultra: one reset if you've paid for under six months, two for six months to a year, three after that. They expire March 8, 2027.
- Pro: one reset, good until November 16.
You only qualify if your paid billing started before launch, November 9 at 10 am Pacific. You'll find your resets, and their expiry dates, in Settings, then Usage.
A reset doesn't raise your limit, and you can't undo one once it's used. You can use it any time before it expires, not only when you've run out. If you cancel or downgrade first, you lose it. On Pro, the sensible time is the day you run a big test and burn through your Fable half.
Fable vs Opus: which one to use for what
Anthropic itself tells most people to start with Opus 5.5 for everyday work. It calls Opus its strongest daily driver and points to Fable for the biggest, most open-ended jobs.
Our own testing backs that up. When we had Fable 5.1, Opus 5.5 and GPT-6 Astra build the same three apps in September, Opus won all three, and the same post has what each build cost. Opus has kept winning our build tests since, including in our Kimi K3.1 review below. Neither test included Fable 5.2, which is why it's worth running your own comparison this week.
So the bigger model hasn't always been the better pick, and 5.2 has something to prove. For a cheaper step down, our five-build Sonnet vs Opus test shows where the smaller models hold up.
How to test Claude Fable 5.2 against Opus this week
Pro subscribers have a rare chance to run the comparison on their own work without paying extra. Pick one big job you'd actually hand to an AI, run it in Opus 5.5, then start a new chat and run the identical prompt in Fable 5.2.
Use a task that genuinely needs the top model, or you won't see a difference. Attach the files in the chat, or point Claude Code at the folder. Two that would:
"Here are our last three years of customer contracts and invoices. Find every place where what we billed doesn't match what the contract says, explain each one, and rank them by how much money is involved."
"Build me a working web app where my team can log client hours, see each client's monthly total against their budget, and export an invoice as a PDF. Ask me questions first if anything is unclear."
Then compare three things: which answer you'd actually use, how long each took, and how much of your weekly limit each one used in Settings, then Usage. The better answer is only worth it if you can afford to run it every week, so that last number matters as much as the answer.
Comments
Sign in to join the conversation.