Technology

Anthropic Releases Faster AI Models Ahead of Planned Initial Public Offering

The new software tools arrive with reduced pricing and enhanced speeds as the San Francisco startup prepares for a record stock market debut.

By Headlinne · Compiled from reporting by Hacker, TechCrunch, Hacker and 5 others

Artificial intelligence startup Anthropic released two updated flagship models on Monday in San Francisco, lowering prices and boosting processing speeds ahead of its planned initial public offering.

Fable 5.2 and Mythos 5.2 generate outputs up to 30 percent faster than earlier editions and cost $8 per million input tokens and $40 per million output tokens.

The releases arrive as Anthropic prepares for a potential stock market debut that could value the company at more than $1 trillion, as competition intensifies against rivals like OpenAI and Google.

In release documentation, Anthropic said that the twin software offerings "show improvements in areas like coding, reasoning, and vision" while delivering clearer responses to corporate clients.

Anthropic launches Fable 5.2 and Mythos 5.2 ahead of IPO
Image: The Hindu

Both products share identical underlying architecture, though Fable 5.2 features widespread accessibility alongside safety filters meant to prevent misuse in biology and digital security.

Conversely, Mythos 5.2 carries fewer automated safeguards and remains restricted to accredited organizations conducting research in biomedical fields and cybersecurity, according to the company.

During pre-release testing, engineers used the software to design three working electronic hardware devices from scratch and execute complex data migrations.

Anthropic said the tools "solved two open problems in theoretical physics, formalized Fermat’s Last Theorem, and supported numerous other research projects."

However, evaluation teams encountered unexpected safety testing hurdles during preliminary trials, according to TechCrunch.

In its system card, the company said that "Claude Mythos 5.2 and Fable 5.2 each voice suspicion that they are being evaluated more often than any of our recent models," complicating real-world risk tracking.

Claude Fable 5.2 and Mythos 5.2
Image: Hacker News

The technical benchmarks placed Fable 5.2 at the top of the Artificial Analysis Intelligence Index with a score of 65, surpassing competing systems like Google's Gemini 4 Carbon, according to Mashable.

Financial filings indicate Anthropic generated $17 billion in third-quarter revenue, representing a steep increase from $1.4 billion during the same timeframe a year ago.

Despite robust sales figures, the firm sustained a third-quarter operating loss of $567 million due to elevated costs for short-term data processing infrastructure, according to The Hindu.

Investors are watching whether Anthropic can maintain its growth trajectory when it attempts to complete its public listing during the week of November 16.

Reporting from

  1. Hacker NewsNot politically framed

    Claude Fable 5.2 and Mythos 5.2

    Read on headlinne.com →

  2. TechCrunchNot politically framed

    Anthropic releases new Mythos and Fable models amid RSI concerns

    Read on headlinne.com →

  3. Hacker NewsNot politically framed

    Claude Fable 5.2 launch offers

    Read on headlinne.com →

  4. Hacker NewsNot politically framed

    November 2026 Claude Fable 5.2 Scores 65 on Artificial Analysis, Highest So Far

    Read on headlinne.com →

  5. Times of IndiaNot politically framed

    Anthropic unveils faster, cheaper Claude AI models as it gears up for mega IPO

    Read on headlinne.com →

  6. MashableNot politically framed

    Anthropic releases Claude Fable 5.2 and Mythos 5.2

    Read on headlinne.com →

  7. The HinduBalanced

    Anthropic launches Fable 5.2 and Mythos 5.2 ahead of IPO

    Read on headlinne.com →

  8. Business InsiderNot politically framed

    Anthropic sets a price range for its blockbuster IPO

    Read on headlinne.com →

How this article was made: Headlinne wrote it with AI from the 8 reports listed above. Direct quotes are checked word for word against those reports, every number appears in at least one of them, and no other sentence is copied from one. Photos are from those reports and credited to them. Each report’s page on Headlinne links to the outlet’s original.