Daily · Article 02 · 5 min read
- Anthropic
Anthropic launches Claude Fable 5.2 and Mythos 5.2
Anthropic notice dated November 9 · reported November 10 JST
Summary
Claude Fable 5.2 and Mythos 5.2 have different access rules. Compare the Claude app’s plan conditions and the 20% lower standard API rates on the same task.
Key facts
Anthropic announced Claude Fable 5.2 and Mythos 5.2 in a release dated November 9. The launch was already being reported early on November 10 in Japan. The two versions share an underlying model. Fable is available on paid Claude plans, subject to plan and organization settings. Mythos access is restricted to eligible organizations. Anthropic announcement TechCrunch
The standard API rates for both versions fell by 20% from 5.1. The new rates are $8 per million input tokens and $40 per million output tokens. What you can use within a Claude app subscription is governed by the plan's own usage rules. Claude Platform release notes
1. Same foundation, different access
Think of a team asking an AI to compare a product manual with a support guide and point to passages that disagree. The model is the system that reads the documents and writes the answer. The app and account determine which version the team can select. Anthropic says the two versions have the same foundation, with stronger safeguards on Fable in sensitive areas. For an ordinary document comparison, you can start from the models available in your Claude account. Anthropic announcement
| Version | What it shares | How access is offered |
|---|---|---|
| Fable 5.2 | Same underlying model as Mythos 5.2 | Paid Claude app plans, subject to plan and organization settings. Also offered through the Claude API |
| Mythos 5.2 | Same underlying model as Fable 5.2 | Restricted organizational access. Anthropic separately lists Claude Security for eligible Enterprise customers |
- Version
Fable 5.2
- What it shares
- Same underlying model as Mythos 5.2
- How access is offered
- Paid Claude app plans, subject to plan and organization settings. Also offered through the Claude API
- Version
Mythos 5.2
- What it shares
- Same underlying model as Fable 5.2
- How access is offered
- Restricted organizational access. Anthropic separately lists Claude Security for eligible Enterprise customers
The cat asks Q&A
The researcherIf our company pays for Claude, can I select Mythos in a regular chat?
Pocket NotebookCheck the model picker for Fable. Mythos has separate access conditions, so a paid account alone does not establish that your team can use it. The Enterprise security product is a separate surface from a regular Claude chat.
2. Two ways to compare documents
A person can compare the documents in Claude, or a company program can request the same work through the API. The route determines which usage rules apply.
%% Two ways to ask Claude to compare the same documents
flowchart TD
B["Person: Claude app"] --> D["Weekly allowance<br/>and usage credits"]
D ~~~ C
C["Program: API"] --> E["Input and output rates<br/>per million tokens"]The cat asks Q&A
- The researcher
Do the lower API rates change what I get in the app?
Pocket NotebookThe app follows your plan’s allowance and credit rules. API requests use the API rates. Anthropic also lists standard API rates for usage-based Enterprise.
3. What your Claude plan includes
A weekly usage limit is the amount of work the plan includes over a week. A seat is an account assigned to one person in a team. Paid plans can offer Fable either within their usage limits or through usage credits, which are a separate way to pay for extra work. On Team and Enterprise, an administrator can control which models people can select. Fable by plan
On web, desktop and mobile, select Fable 5.2 from the model picker. Enterprise admins can enable it earlier. Seven days after launch, it is added to roles with an existing Fable model enabled unless an admin opts out beforehand.
The launch offer includes all Fable models in the eligible plans. For Pro and standard Team seats, after the offer ends or the Fable share is used up, further Fable usage is billed to usage credits if they are enabled. The account must meet the separate eligibility conditions to receive a promotional limit reset. Launch offer conditions
The original offer deadline for Pro and standard Team seats is November 16, 2026 at 11:59 p.m. PT.
| Claude app plan or seat | Regular usage | Launch offer |
|---|---|---|
| Free | Fable is unavailable | Ineligible |
| Pro or standard Team seat | Billed to usage credits | Up to 50% of the existing weekly allowance until November 17 at 4:59 p.m. JST |
| Max, Ultra, premium Team or premium seat-based Enterprise | Up to 50% of the existing weekly allowance for Fable | Fable limit resets for eligible accounts |
| Standard seat-based Enterprise | Available if the organization enables usage credits | Ineligible |
| Usage-based Enterprise | Billed at standard API rates. The organization controls availability | Ineligible |
- Claude app plan or seat
Free
- Regular usage
- Fable is unavailable
- Launch offer
- Ineligible
- Claude app plan or seat
Pro or standard Team seat
- Regular usage
- Billed to usage credits
- Launch offer
- Up to 50% of the existing weekly allowance until November 17 at 4:59 p.m. JST
- Claude app plan or seat
Max, Ultra, premium Team or premium seat-based Enterprise
- Regular usage
- Up to 50% of the existing weekly allowance for Fable
- Launch offer
- Fable limit resets for eligible accounts
- Claude app plan or seat
Standard seat-based Enterprise
- Regular usage
- Available if the organization enables usage credits
- Launch offer
- Ineligible
- Claude app plan or seat
Usage-based Enterprise
- Regular usage
- Billed at standard API rates. The organization controls availability
- Launch offer
- Ineligible
The cat asks Q&A
- The researcher
Does “up to 50%” add half a week's usage to my account?
Pocket NotebookIt is a share of the weekly allowance you already have. Usage of other Claude models also draws on the overall allowance. Check
Settings → Usagefor what remains, and ask your Team or Enterprise admin if Fable does not appear in the picker.
4. The price of the same API job
The API lets another program send the documents to Claude and receive an answer. It charges based on tokens, the units used to count pieces of text and other content. Input is what the program sends, and output is what Claude generates. The table shows Anthropic’s published rates in US dollars per million tokens. Claude Platform release notes
In a fictional API job with 100,000 input tokens and 25,000 output tokens, the standard 5.1 cost is 0.1 × $10 + 0.025 × $50 = $2.25. At standard 5.2 rates the same quantities cost $1.80, a saving of $0.45 or 20%. Fast mode would cost $3.60. This example uses the listed token rates and leaves out caching, separately priced tools and taxes. Anthropic says fast mode can generate tokens up to 2.5 times as fast on supported routes. Document loading and other work still contribute to the time you wait. Anthropic announcement
| API model and mode | Input per million | Output per million |
|---|---|---|
| Fable or Mythos 5.1, standard | $10 | $50 |
| Fable or Mythos 5.2, standard | $8 | $40 |
| Fable or Mythos 5.2, optional fast mode | $16 | $80 |
- API model and mode
Fable or Mythos 5.1, standard
- Input per million
- $10
- Output per million
- $50
- API model and mode
Fable or Mythos 5.2, standard
- Input per million
- $8
- Output per million
- $40
- API model and mode
Fable or Mythos 5.2, optional fast mode
- Input per million
- $16
- Output per million
- $80
The cat asks Q&A
- The researcher
If my team switches, will our total API bill fall by 20%?
Pocket NotebookThe 20% applies when you compare the standard rates with the same input and output quantities. If the new model writes longer answers or you run more jobs, your total usage changes. Compare the usage records for the same job before estimating your bill.
To close
Check the route you actually use
For a person working in Claude, the model picker and usage screen show what the account can use. For a team connecting a service through the API, compare the input and output counts from the same task before deciding whether the change saves money.
Sources
- Anthropic’s launch announcementanthropic.com
- TechCrunchtechcrunch.com
- Claude Platform release notesplatform.claude.com
- Claude Help Center plan rulessupport.claude.com
- Claude Help Center launch offersupport.claude.com