Anthropic's Claude Fable 5 Surpasses GPT 5.5 with Landmark Performance

Claude Fable 5 Unveiled

According to НВ — Техно: On June 9, 2026, Anthropic launched its latest AI model, Claude Fable 5, alongside a restricted version called Mythos 5. Accessible through the Claude API and enterprise plans, the model costs $10 per million input tokens and $50 per million output tokens. Until June 22, subscribers on Pro, Max, Team, and Enterprise plans can use Fable 5 at no extra charge, but starting June 23, Anthropic will introduce a credit-based system for usage.

Capabilities and Benchmarks

Anthropic noted that in high-risk areas-covering biology, chemistry, and cybersecurity-the model blocks responses and automatically switches to Claude Opus 4.8. The company asserts that Fable 5 and Mythos 5 outperform competitors like GPT 5.5 and Gemini 3.1 Pro in coding tasks. Analytics firm Hex confirmed that Fable 5 is the first model to achieve a 90% score on their baseline test.

During Fable 5's development, Anthropic ran an external bug-hunting program lasting over 1,000 hours, which uncovered no universal jailbreaks.

“We conducted an external bug-hunting program that found no universal jailbreaks after more than 1,000 hours of testing. We then worked with external organizations, which also failed to find any universal jailbreaks,” the company stated in its release.

Additionally, Fable 5 demonstrated high effectiveness on complex questions, showing strong judgment and attention to nuance.

“On the most challenging questions, [the model] shows strong judgment and attention to nuance,” noted analysts at Hex.

Reports indicate that the Base44 platform highlighted Fable 5's efficiency in rapidly building entire applications, while Genspark cited significant advantages in design and game code generation. Rakuten's service also pointed out the system's unique ability to self-check its own work.

The launch of Claude Fable 5 underscores the intensifying competition in the AI market, as companies continuously refine their models to deliver superior results across various domains. Given the high marks from analysts and positive user feedback, Fable 5 could become a key tool for enterprises seeking efficient solutions for automation and complex application development. The pricing changes set for June 23 may influence subscriber numbers and the broader adoption of this new technology in business settings.

As Anthropic continues to innovate with its AI models, the recent introduction of a restricted access version for vulnerability discovery highlights the company's commitment to enhancing security measures in AI technology. This initiative complements the impressive capabilities of Claude Fable 5, showcasing how advancements in AI can also play a crucial role in identifying and mitigating risks. For more details on this new development, check out our coverage of Anthropic's latest AI model focused on vulnerability detection.