Anthropic Launches Claude Fable 5.2 and Mythos
Anthropic has introduced two new AI models, Claude Fable 5.2 and Claude Mythos 5.2, strengthening its AI portfolio ahead of a potential initial public offering (IPO).
Announced on 9 November 2026, the releases target demanding reasoning, complex problem-solving and long-running autonomous tasks. The launch also reinforces Anthropic’s efforts to compete in the rapidly expanding enterprise AI market.
The two models share the same underlying capabilities. However, Anthropic has set different access requirements for each version, reserving Mythos for organisations that meet its verification requirements.
The launch also comes as Anthropic prepares for a potential stock market listing. Consequently, its latest releases carry significance for both its technology strategy and its commercial ambitions.
Claude Fable 5.2 Targets Advanced Reasoning and AI Agents
Anthropic describes Claude Fable 5.2 as a model designed for demanding reasoning and long-horizon agentic work.
In practical terms, this positioning targets tasks that require sustained analysis, multiple reasoning stages and coordinated actions. As a result, the model aims to support more complex workflows across enterprise applications and AI-powered agents.
The new model includes a one-million-token context window and supports up to 128,000 output tokens. It also features always-on adaptive thinking, allowing the model to adjust its reasoning to the task.
Anthropic has also made Fable 5.2 available across several platforms:
-
Claude API and Claude Platform.
-
Amazon Web Services through Anthropic-managed infrastructure.
-
Google Cloud through Anthropic-managed and Google-managed services.
-
Microsoft Foundry.
The company has also introduced a revised pricing structure for the new models. Claude Fable 5.2 costs $8 per million input tokens and $40 per million output tokens.
Claude Mythos 5.2 Introduces Restricted AI Access
While Fable targets broader enterprise and developer adoption, Mythos follows a more restricted distribution model.
Anthropic limits Claude Mythos 5.2 to organisations verified through its security programmes. These include participants in the company’s Cyber Verification Program.
This restriction reflects the model’s intended use in specialised environments. In particular, approved organisations can access its advanced capabilities for security-related work.
Both models feature a one-million-token context window and support up to 128,000 output tokens. Additionally, they include adaptive thinking capabilities designed to handle demanding tasks.
Anthropic has also reduced the models’ standard API pricing compared with their previous versions. The new pricing stands at:
| Pricing | Claude Fable 5.2 and Mythos 5.2 |
|---|---|
| Input tokens | $8 per million |
| Output tokens | $40 per million |
The company has also introduced a fast mode for Fable 5.2 through its Messages API. This option delivers up to 2.5 times faster output, though it carries higher usage charges.
IPO Plans Add Commercial Significance to Model Launch
The new releases arrive as Anthropic prepares for its anticipated public market debut.
The company has already filed its IPO registration statement and is preparing for a potential November 2026 listing. Its planned Nasdaq ticker is ANT.
However, the offering faces uncertainty over pricing and investor expectations. On 10 November, reports emerged that banks circulated preliminary offering details before Anthropic withdrew them as inaccurate.
The company is therefore navigating significant commercial expectations alongside its expanding AI development programme.
The timing also reflects the growing financial demands of developing advanced AI systems. Anthropic continues to invest in computing infrastructure, model research and enterprise distribution.
Meanwhile, competition with OpenAI and other AI developers continues to intensify. Consequently, the launch of Fable 5.2 and Mythos 5.2 strengthens Anthropic’s product portfolio ahead of a potentially significant transition.
The two models also demonstrate the company’s focus on specialised AI applications. Fable targets broader reasoning and agentic workflows, while Mythos provides restricted access for verified organisations.
