- Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on Monday, Nov. 9, 2026.
- Pricing is $8 per million input tokens and $40 per million output tokens, 20% below the 5.1 generation.
- Output is up to 30% faster, and an optional fast mode raises token generation speed by up to 2.5 times.
- Fable 5.2 is on all platforms, while Mythos 5.2 is limited to organizations that pass verification.
- The benchmark scores are Anthropic's own, and Gemini 4 Carbon still leads one business workflow test.
Anthropic has released Claude Fable 5.2 and Claude Mythos 5.2 at a price 20% below the previous generation. The launch comes as the company prepares for a U.S. listing, and it is Anthropic's first frontier release since CEO Dario Amodei called for a slower pace of AI development in September.
In its announcement, Anthropic said both models improve on coding, reasoning and vision. Fable and Mythos share the same underlying model. The difference is that Fable carries strict cybersecurity and biology safeguards, which is what allows the company to offer it to all users.
Quick answer: Claude Fable 5.2 and Mythos 5.2 launched on Nov. 9, 2026, at $8 per million input tokens and $40 per million output tokens, 20% below the 5.1 generation. Fable 5.2 is available on all platforms, including AWS, Google Cloud and Microsoft Azure. Mythos 5.2 is limited to cybersecurity and biomedical organizations that pass a verification program.Claude Fable 5.2 Pricing Compared With Rivals
The price cut is the change developers will notice first. CNBC reported that the new rate is 20% below the 5.1 generation, while Fortune noted it is also 20% below OpenAI's GPT-6 Astra for ordinary input and output.
| Model | Input per 1M tokens | Output per 1M tokens | Note |
|---|---|---|---|
| Claude Fable 5.2 | $8 | $40 | Standard |
| Claude Fable 5.2 fast mode | $16 | $80 | Up to 2.5x faster |
| Earlier Fable generation | $10 | $50 | Fable 5 launch price in June |
| GPT-6 Astra | $10 | $50 | According to Fortune |
Prices in U.S. dollars. Fable 5.2 rates are from Anthropic's announcement.
As a rough example, a request with 100,000 input tokens and 20,000 output tokens now costs about $1.60, down from $2 at the old rate. That gap adds up for companies running thousands of tasks a day.
Test Results: Ahead on Most Benchmarks, Not All
Anthropic published a comparison table covering its own models and its main competitors. All figures below come from Anthropic, so they should be read as company claims until independent testing is available.
| Benchmark | Fable 5.2 | Opus 5.5 | GPT-6 Astra | Gemini 4 Carbon |
|---|---|---|---|---|
| FrontierCode 2.0 (coding) | 49.1% | 45.5% | 42.8% | 41.1% |
| FrontierSWE v2 (coding) | 69.9% | 62.3% | 65.5% | 61.2% |
| AutomationBench (business workflows) | 51.4% | 42.5% | 41.4% | 53.9% |
| Terminal-Bench-Science (research) | 74.6% | 58.7% | 68.1% | 64.8% |
| OSWorld 2.1 (computer use) | 89.3% | 87.2% | 79.5% | 81.5% |
| Chartography (reading charts) | 73.4% | 64.4% | 71.0% | 72.5% |
Source: Anthropic announcement, Nov. 9, 2026. Effort settings can differ by model.
Fable 5.2 does not win every category. On AutomationBench, which measures business workflows, Gemini 4 Carbon scored 53.9% against 51.4% for Fable 5.2. On Chartography, its lead over Gemini 4 Carbon is only 0.9 percentage points.
An Operating System, Hardware and Fermat's Last Theorem
Anthropic highlighted several experiments meant to show long-running work. In one, an engineer asked Fable 5.2 to design a prototype operating system from an empty folder. The model coordinated more than 1,000 subagents over eight days and produced a system that boots on a laptop, with drivers, file storage, windows, a login screen and an installer.
On hardware, Fable 5.2 was given a one-paragraph prompt to create three devices: a home voice-assistant hub, a USB speakerphone and a handheld virtual pet. Working with circuit board company Diode, all three functioned without electrical modifications.
In mathematics, a team of 1,000 Fable 5.2 agents formalized the proof of Fermat's Last Theorem in 30 hours with minimal human intervention. Anthropic said Opus 5.5, in the same setup, completed only 26% in 18 hours. The company also said Fable 5.2 solved two open problems in black hole physics. Research claims like these still need to be assessed by the scientific community.
Who Can Use Fable 5.2 and Mythos 5.2
Fable 5.2 is available on all platforms, including Amazon Web Services, Google Cloud and Microsoft Azure, under the model name claude-fable-5-2 on the Claude Platform. Eligible paid subscribers get extra usage this week in the form of one to three Fable limit resets. Fable can also take up to 50% of weekly usage limits on Pro plans and standard Team seats.
Mythos 5.2 is not open to the public. Organizations in cybersecurity and biomedical research must apply through the Cyber Verification Program or the Life Sciences Verification Program.
Anthropic does not recommend Fable for everything. For most coding and knowledge work, the company suggests starting with Claude Opus 5.5, then moving up to Fable or Mythos when the job is especially demanding and open-ended.
Fewer Blocks on Legitimate Requests
A common complaint about heavily safeguarded models is that they refuse reasonable requests. Based on early testing across claude.ai, Claude Code and Cowork, Anthropic reported about 55% fewer blocks with Fable 5.2 than with Fable 5.1. That breaks down to about 60% fewer in cybersecurity and 40% fewer in biology.
The company said high-stakes risks from misalignment remain low. Before release, the underlying model was also tested by outside evaluators, including the UK AI Security Institute, Frontier Design and METR. Summaries of their reports appear in the system card.
A Launch Ahead of the IPO and After a Call to Slow Down
The timing has drawn market attention. CNBC reported that Anthropic's IPO could take place as soon as next week, with a target valuation of $1 trillion to $1.5 trillion. That is below the $2 trillion figure floated in recent months.
The release has also raised questions because Amodei argued in September that AI development should slow down. Dianne Penn, Anthropic's head of product management for research, gave CNBC a short answer: "We called for pacing, not pausing."
FAQ (Frequently Asked Questions)
It costs $8 per million input tokens and $40 per million output tokens. Fast mode costs $16 and $80. They share the same underlying model. Fable has strict cybersecurity and biology safeguards and is open to all users, while Mythos is limited to verified organizations. In Anthropic's table, Fable 5.2 leads on all six tests listed in this article. Those numbers come from the company, so independent results are still worth waiting for. Yes, on paid plans that include Fable models. On Pro plans and standard Team seats, Fable can use up to 50% of weekly limits. CNBC reported it could happen as soon as next week. The model announcement did not give a date.1. How much does Claude Fable 5.2 cost?
2. What is the difference between Claude Fable 5.2 and Mythos 5.2?
3. Is Fable 5.2 better than GPT-6 Astra?
4. Can regular subscribers use Fable 5.2?
5. When is the Anthropic IPO?
Conclusion
Fable 5.2 pairs two things enterprise buyers want, higher scores and a lower price. Most of the performance evidence, however, still comes from Anthropic itself.
The practical approach is the one the company recommends: start with a daily-use model, then move up to Fable for long jobs that cheaper models cannot handle.
- Fable 5.2 and Mythos 5.2 launched on Nov. 9, 2026.
- Prices fell 20% to $8 input and $40 output per million tokens.
- Mythos 5.2 is available only through verification programs.
- Blocked requests on Fable fell about 55% compared with version 5.1.
- Published scores come from Anthropic and have not been independently tested.
