In a notable development for the AI sector, Anthropic has launched Claude Fable 5, a general-access version of its advanced Mythos-class AI. This release signifies an evolution in the company’s offerings, claiming to outperform all previously available models. The introduction comes at a time when interest in AI applications is surging across various industries, especially in areas that require stable analytical capabilities and cybersecurity measures.
Performance Highlights
Anthropic presents Claude Fable 5 as a leading performer on various capability benchmarks, particularly excelling in longer and more complex tasks. Early-access results have highlighted its effectiveness in real-world applications. For example, Stripe reported that Fable 5 managed to migrate a 50 million line Ruby codebase in a single day—an effort that typically demands a full team more than two months to complete manually.
Further evaluations by Cognition’s FrontierCode revealed that Fable 5 produced code meeting production-quality standards more effectively than its competitors, even under medium effort conditions. In the financial sector, benchmarks from Hebbia indicated that Fable 5 ranked first for senior-level analytical reasoning, showing marked improvements in document reasoning and chart interpretation. IMC, a trading firm, confirmed these findings, pointing out the model's exceptional skills in analyzing expected values and conducting fact-checking.
Advancements in Gaming and Cybersecurity
The model's capabilities reach beyond traditional analytical tasks. Anthropic claims that Fable 5 completed the game Pokémon FireRed using only raw screenshots, a task that earlier versions required additional tools to perform. This achievement demonstrates the model's advanced understanding of visual data processing.
In cybersecurity, the restricted variant, Claude Mythos 5, is being provided to US government cyber defenders through the Project Glasswing program. Anthropic asserts that Mythos 5 has the strongest cybersecurity capabilities of any model globally, although this claim remains self-assessed and lacks independent verification.
Pricing and Market Implications
Both Claude Fable 5 and Mythos 5 were launched on June 9, 2026, with competitive pricing set at $10 per million input tokens and $50 per million output tokens for the restricted version. This pricing strategy marks a significant reduction compared to fees for the earlier Claude Mythos Preview, potentially making advanced AI functionalities more accessible to a wider audience.
The launch of Claude Fable 5 represents a new chapter in Anthropic’s trajectory, pushing the limits of what AI can achieve in business and security contexts. As enterprises and government bodies increasingly depend on AI to enhance efficiency and security, the implications of this release may resonate across various sectors, influencing the future of intelligent systems.
The stories that move AI & crypto markets — before the market reacts.
Free. 7am ET. Five stories. 62,400 readers.
