AIToday
Large Language ModelsAI Business & IndustryThe Verge AIPublished: Sep 2, 2026, 10:00 JST2 min read

Anthropic launches cheaper Claude Fable 5.1

Anthropic launches cheaper Claude Fable 5.1

Key takeaway

  • Anthropic launched two new AI models. Fable 5.1 is cheaper and stronger than its predecessor.

  • It also has more precise safeguards and better data retention options.

  • Mythos 5.1 is available only to Project Glasswing participants.

3 Key Points

  1. What happened

    Anthropic has launched two new AI models, Fable 5.1 and Mythos 5.1. Fable 5.1 is about 25 percent cheaper typically and up to 45 percent cheaper for complex agentic tasks, while offering stronger performance than Fable 5.

  2. Why it matters

    The new models address customer criticisms about price, data retention, and overzealous safeguards. Early users like Every CEO Dan Shipper praise the model's speed and natural language, while Box CEO Aaron Levie says its agent picked up on subtleties in data that Fable 5 missed.

  3. What to watch

    Fable 5.1 is available on all platforms, while Mythos 5.1 is limited to Project Glasswing participants. Enterprise Frontier Safeguards, which store data on customer cloud servers, will roll out later this fall.

Ask the AI about this article →

Context & Analysis

The launch of Fable 5.1 and Mythos 5.1 is a direct response to customer feedback about price, data retention, and overly strict safety filters. The lower pricing on cached data processes addresses cost complaints without sacrificing performance, which early impressions from notable figures like Dan Shipper and Aaron Levie suggest is improved. The more precise safeguards aim to reduce unnecessary blocks on legitimate queries, while still maintaining restrictions on sensitive topics like biology, and cybersecurity tasks involving exploit generation are still routed to Opus models.

Anthropic is also addressing data privacy with Enterprise Frontier Safeguards, which store data on the customer's cloud servers instead of its own, offering complete privacy. This feature is scheduled to roll out later this fall, which may appeal to enterprise customers with strict data handling requirements. The company's statement that Fable 5.1 is now allowed for identifying software vulnerabilities while other cybersecurity tasks use Opus models shows a nuanced approach to safety, though the practical implications remain to be seen.

FAQ

How much cheaper is Claude Fable 5.1?
Fable 5.1 costs about 25 percent less typically and up to 45 percent less for complex agentic tasks compared with Fable 5.
What are the new data retention options?
Enterprise Frontier Safeguards offer complete privacy by storing data on the customer's cloud servers instead of Anthropic's own. They will start rolling out later this fall.
Which tasks still use Opus models?
Anthropic will still redirect some cybersecurity tasks to Opus models, such as penetration testing, exploit generation, and binary-based vulnerability scanning.

Also reported by ITmedia AI+, THE DECODER

Get the latest Large Language Models news every morning

For example, today's edition would include:

  • LLM serving: why continuous batching winsDaily Dose of Data Science · 2h ago
  • Anthropic's Claude Fable 5.1 Now on Snowflake Cortex AISnowflake AI Blog · 2h ago
  • BenchMIRT: AI benchmarks may measure the wrong skillsHugging Face Blog · 2h ago

AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.

Free · takes 30 seconds · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. Q&As are published on this page for other readers too.

Related Articles

Next articleOpenAI's Astra model nears release, excels at hacking