Anthropic released Claude Fable 5.2 on November 9, 2026, alongside Claude Mythos 5.2. Fable 5.2 is the most capable Claude model open to the public. It costs $8 per million input tokens and $40 per million output tokens (20% less than Fable 5.1), generates output up to 30% faster and adds an optional fast mode. Within three days, developers, creatives and educators had published games, animated videos and working electronic devices built with it.
If you want to move from reading about these models to using them at work, the Founderz Online Program in AI & Innovation teaches you to work with Claude, ChatGPT and Microsoft 365 Copilot, and to build your own AI agents. The frameworks you learn work across every AI platform, so they keep paying off as new models like Fable 5.2 arrive. The program is developed in collaboration with Microsoft.
What is Claude Fable 5.2, and how does it differ from Mythos 5.2?
Claude Fable 5.2 and Claude Mythos 5.2 share the same underlying model. The difference is the safeguards. Fable runs classifiers for cybersecurity and biology, which is why Anthropic can offer it to every paying user. Mythos 5.2 is limited to organizations vetted through Anthropic’s Cyber Verification and Life Sciences Verification programs, and those verification programs are the only route to Mythos access.
Fable sits above Opus in the Claude lineup, but Anthropic still recommends Opus 5.5 as the starting point for most coding and knowledge work. Anthropic says Fable 5.2 is designed for the most ambitious, open-ended technical work. It can also act as an advisor that Claude Haiku 5.5, Sonnet or Opus consult on hard problems, or as an orchestrator that hands simpler tasks to smaller models.
For context on how far the Claude family has come, our earlier piece on Claude 3 and its business uses covers the generation that started Anthropic’s push into business work.
Fable 5.2 vs Fable 5.1: what’s new
Anthropic reports improvements in coding, reasoning and vision. These are the changes users will notice first:
- Lower API price: $8 / $40 per million input and output tokens, down from $10 / $50 on Fable 5.1.
- More speed: Fable 5.2 generates output up to 30% faster than Fable 5.1.
- Optional fast mode: token generation up to 2.5 times faster, at $16 / $80 per million.
- Fewer unnecessary blocks: in early testing across claude.ai, Claude Code and Cowork, Anthropic reports fewer blocks than with Fable 5.1, about 55% fewer overall.
- Clearer communication: testers say the model gives more direct answers and admits when it doesn’t know.
The context window stays at 1 million tokens, with a 128K-token output ceiling and a June 2026 knowledge cutoff. Older versions remain available: Fable 5.1 and the original Fable 5 are still on paid plans, so Claude is keeping Fable 5 for now.
Fable 5.2 pricing and API price
Fable 5.2 costs $8 per million input tokens and $40 per million output tokens on the Claude API, and every related rate fell by the same share. Figures checked on November 12, 2026 against Anthropic’s official documentation:
| Rate (per million tokens) | Fable 5.2 | Fable 5.1 |
|---|---|---|
| Input | $8 | $10 |
| Output | $40 | $50 |
| Cache read | $0.20 | $0.25 |
| Fast mode (input / output) | $16 / $80 | Not offered |
A lower price per token does not guarantee a lower bill. Fable 5.2 has five effort levels, and the longer the model thinks, the more tokens it uses. On FrontierCode 2.0, Anthropic’s system card puts the average cost per task at $2.10 on low effort and $44.70 on max.
On a Claude plan, Fable 5.2 is not available for free. Max, Ultra and premium Team and Enterprise seats include Fable usage for up to 50% of weekly limits. Pro and standard seats pay through usage credits, except during the launch offer: from November 9 to 16, 2026, they also include Fable for up to 50% of their weekly limit.
Fable 5.2 is available on every major platform
- Claude apps and Claude Code on paid Claude plans, with the 1M-token window by default in Claude Code.
- Claude Platform, the Claude API, as
claude-fable-5-2. - Cloud: Amazon Web Services (Bedrock), Google Cloud and Microsoft Azure through Microsoft Foundry.
Fast mode works in Claude Code, the API, Google Cloud and Microsoft Foundry. For companies that work with Microsoft, Foundry support means they can build AI agent workflows with Fable inside the same governance and monitoring layer they already use for Azure. Claude Enterprise customers can also use Claude Security, which now runs on Mythos 5.2.
Fable 5.2 benchmark results and coding performance
Anthropic says Fable 5.2 and Mythos 5.2 are a step above Opus 5.5 in almost every capability it tested. These are vendor-reported results, with each model at its highest available effort:
| Benchmark (Anthropic data, Nov 9, 2026) | Fable 5.2 | Fable 5.1 |
|---|---|---|
| FrontierCode 2.0 (agentic coding) | 49.1% | 32.4% |
| Humanity’s Last Exam (with tools) | 74.2% | 65.6% |
| Remote Labor Index (knowledge work) | 41.3% | 17.9% |
| Chartography (chart reading) | 73.4% | 44.8% |
| OSWorld 2.1 (computer use, partial score) | 89.3% | 83.5% |
Fable 5.2 scored 49.1% on FrontierCode 2.0, the highest of any released model Anthropic tested. It loses on two fronts: Gemini 4 Carbon leads AutomationBench, and Opus 5.5 scores slightly higher on the AA-Briefcase Elo rating. The most useful figure for budgets: at high effort, Fable 5.2 matches Opus 5.5’s best FrontierCode score for roughly half the cost per task.
On coding, Anthropic says over 90% of the changes merged into its core codebase are now written by Claude with human guidance. In one test, Fable 5.2 coordinated more than 1,000 subagents for eight days to build an operating system that boots on a laptop.
The creative explosion Fable 5.2 started this week
Fable 5.2 had been out for a few hours when the first projects started circulating, and what people shared most were things you can play, watch and hold.
Whole games from one prompt. Felix Rieseberg, who works at Anthropic, posted Saltline, a first-person shooter set in a rain-soaked port. One prompt produced it, with no human feedback during the run, in about eight hours and for about $550 at API prices. Creative coder Kevin Ngo posted What the Water Keeps, a browser-playable boat ride through eight dreamlike worlds, and was upfront that he used other AI tools for the 3D models and music.
Video and animation. Wharton professor Ethan Mollick asked Fable 5.2 to build the most impressive thing it could for his blog and got back a 90-second video with its own soundtrack. Other creators turned a tweet and a brand book into an animated video, and a 49-second voiceover into a 2D animated short.
Real hardware and science. From a one-paragraph prompt, Fable 5.2 designed three working electronic devices: a home voice-assistant hub, a USB speakerphone and a handheld virtual pet. A team of 1,000 agents formalized the proof of Fermat’s Last Theorem in 30 hours, and the model solved two open problems in the physics of black holes.
Is Fable 5.2 worth it?
The viral demos come with fine print. Many came from people with early access or from companies that sell AI tools, and several users say the model needs long, detailed prompts. Some Pro subscribers burned through much of their five-hour limit in under an hour. The $550 game is a useful reminder: a prompt can be short, but the work behind it is not free.
For everyday work, Anthropic still recommends Opus 5.5 as the starting point. Fable 5.2 earns its place on big jobs: software migrations, reviews of entire document sets, multi-hour research or prototypes that used to need a team. Getting value from any new model takes knowing which one fits each task, how to write precise instructions, how to design agents and how to weigh cost against results. The Online Program in AI & Innovation covers exactly that, 100% online and at your own pace.
Claude Fable 5.2 FAQs
When was Claude Fable 5.2 released?
Anthropic released Claude Fable 5.2 on November 9, 2026, together with Claude Mythos 5.2. It replaces Fable 5.1, released on September 1, 2026, as Anthropic’s most capable publicly available model. Anthropic says it will not retire Fable 5.2 before November 9, 2027.
Can Claude Fable 5.2 be used for free?
No. Fable 5.2 needs a paid Claude plan or API access. During launch week, from November 9 to 16, 2026, Pro users and standard Team seats can use Fable models at no extra cost for up to 50% of their weekly limit, and each gets one Fable usage reset.
Can you use Claude Fable 5.2 on Microsoft Azure?
Yes. Fable 5.2 has been available on Microsoft Azure through Microsoft Foundry since launch day, as well as on Amazon Web Services, Google Cloud and the Claude API. Fast mode also works in Foundry, so companies that work with Microsoft can build Fable agents inside their Azure environment.
Should I use Fable 5.2 or Opus 5.5?
Fable 5.2 performs better on long, complex, open-ended tasks and costs twice as much per token as Opus 5.5. Anthropic recommends Opus 5.5 as the everyday model and Fable for the most ambitious technical work, where the higher cost per token pays off in work other models cannot finish.
Where can I learn to use AI models like Claude at work?
The Founderz Online Program in AI & Innovation teaches you to use Claude, ChatGPT and Microsoft 365 Copilot on real tasks, build your own AI agents and apply AI responsibly, with frameworks that work across every AI platform. It is 100% online and self-paced, developed in collaboration with Microsoft, and assessed through practical challenges you can apply to your job.
Pricing, availability and benchmark data checked on November 12, 2026 against Anthropic’s official announcement, Claude Platform documentation and the Claude Help Center. Benchmark results are vendor-reported.
