Claude Fable 5.2 pricing dropped to $8 per million input tokens and $40 per million output tokens, down 20% from Fable 5.1. But a cheaper token doesn’t automatically mean a cheaper job. Early third-party cost testing suggests that at peak effort, a single task can cost about twice as much as it did on 5.1.
Anthropic announced Fable 5.2 and Mythos 5.2 on November 9, 2026. Here’s what changed, what it really costs, and when you should skip it.
- Released November 9, 2026, as the second major update to the Fable 5 class since June.
- List price: $8 input / $40 output per million tokens (was $10 / $50).
- Outputs run up to 30% faster than Fable 5.1, with gains in coding, reasoning and vision.
- Mythos 5.2, the restricted sibling, is limited to vetted organizations.
What changed in Claude Fable 5.2 pricing versus 5.1
The headline number is simple: both input and output prices fell by 20%. Speed went the other direction, in your favor, with outputs up to 30% faster than before. SD Times reports that Fable 5.2 is available on all platforms from day one and shows stronger engineering judgment on ambiguous requests.
The table below puts the two versions side by side. The per-task figure comes from Artificial Analysis testing at peak effort, as summarized by Stack Archive.
| Metric | Fable 5.1 | Fable 5.2 |
|---|---|---|
| Input price (per 1M tokens) | $10 | $8 |
| Output price (per 1M tokens) | $50 | $40 |
| Output speed | Baseline | Up to 30% faster |
| Cost per task at peak effort | $12.99 | $26.50 |
The first three rows look like a clear win. The last row is the catch, and it’s the one that decides your invoice.
Why cheaper tokens can still mean a bigger bill
Token price is only half of the equation. The other half is how many tokens a model burns to finish the job. Fable 5.2 reportedly used about 98,000 output tokens per task, which is actually fewer than Opus 5.5’s 173,000 at its Max tier.
Even so, the peak-effort cost landed at $26.50 per task, roughly 2.6 times Opus 5.5’s $10. So the fix for your budget isn’t the price list. It’s measuring what a completed task costs on your own workload.
A simple way to do that: log total input and output tokens for ten representative jobs, multiply by the new rates, and compare the totals across models. If a job needs retries or long reasoning chains, those tokens count too, and they’re easy to forget.
This isn’t unique to Anthropic. Google’s Gemini 4 Carbon pricing change and OpenAI’s GPT-6.2 price moves raise the same question: list price rarely tells you the real cost.
Effort level is your real cost dial
The most useful finding is about effort settings. Anthropic says that at High effort, Fable 5.2 scores 45.3% on FrontierCode, close to Opus 5.5’s 45.5%, for roughly half the cost per task.
That’s a vendor claim, so treat it as a starting point rather than a guarantee. Still, it points to a practical routine for anyone calling the Claude API:
- Run a sample of your real tasks at High effort first.
- Compare cost per completed task against Opus 5.5, not just accuracy.
- Reserve peak effort for the hard problems where the accuracy gap pays for itself.
Why this matters: Fable 5.2 leads the Artificial Analysis Intelligence Index at 65 versus 59 for Opus 5.5, but that lead is priciest at the top setting. It also trails Opus 5.5 on agentic knowledge work, per the same summary.
Who should use Fable 5.2, Mythos 5.2 or stay on Opus 5.5
Anthropic itself says Fable and Mythos suit its most ambitious, open-ended technical work. For most coding and knowledge tasks, it recommends starting with Claude Opus 5.5, which it calls its strongest daily driver.
Mythos 5.2 isn’t something you can simply sign up for. Per SD Times, access runs through Anthropic’s Life Sciences and Cyber Verification Programs. If you’re a regular subscriber, Fable 5.2 is your option, and for casual chat or light coding you probably won’t notice much difference from Opus 5.5.
Subscribers also got a bonus: plans that include Fable received one to three usage resets, depending on how long they’ve subscribed, according to 9to5Mac. If you’re curious about Anthropic’s other recent launches, our look at Claude Sheets and the new office tools covers the product side.
Frequently Asked Questions
How much does Claude Fable 5.2 cost?
It costs $8 per million input tokens and $40 per million output tokens. That’s 20% below Fable 5.1’s $10 and $50.
Is Claude Fable 5.2 cheaper than Fable 5.1?
Per token, yes. Per task at peak effort, Artificial Analysis measured $26.50 versus $12.99 for 5.1, so your real cost depends on the effort setting you choose.
Is Claude Fable 5.2 better than Opus 5.5?
It leads on the Artificial Analysis Intelligence Index (65 vs 59) but trails on agentic knowledge work. Anthropic still recommends Opus 5.5 as the default for most work.
Can anyone use Mythos 5.2?
No. It’s limited to vetted organizations through Anthropic’s Life Sciences and Cyber Verification Programs.
The bottom line on Fable 5.2
Don’t switch your production workloads just because the price list got shorter, since the savings only show up when a task stays at a modest effort level. Keep Opus 5.5 as your default, test Fable 5.2 at High effort on a handful of your hardest tasks, and only move over if the cost per finished task holds up. The cut is real, but so is the bill at maximum effort.


Leave a Reply