Anthropic releases Mythos 5.2 and Fable 5.2, saying they are "best for the most ambitious, open-ended technical work", for $8/1M input and $40/1M output tokens

The cost of AI inference decreases significantly for ambitious technical work.

Anthropic has released Claude Fable 5.2 and Claude Mythos 5.2, positioning them as superior for complex technical tasks. These new models reportedly offer performance improvements in coding, reasoning, and vision, exceeding previous versions and even rivaling or surpassing Claude Opus 5.5 on various benchmarks. Specifically, Fable 5.2 and Mythos 5.2 demonstrate faster output generation, up to 30% quicker than their predecessors, and are priced more affordably at $8 per million input tokens and $40 per million output tokens. An optional fast mode is also available for these models, doubling generation speed at a higher cost. Mythos 5.2 is designed for cybersecurity and biomedical research applications, requiring an application process, while Fable 5.2 includes robust safeguards and is available to all users. Anthropic states these models have successfully assisted in complex projects like software migrations, hardware design from scratch, and advancements in theoretical physics research.

Anthropic's new Fable 5.2 and Mythos 5.2 models signal a strategic move towards democratizing access to highly advanced AI capabilities, previously reserved for the most resource-intensive operations. By substantially lowering the price point for demanding tasks, Anthropic is enabling a wider range of organizations and researchers to tackle complex challenges in areas like advanced coding, scientific discovery, and engineering that were previously cost-prohibitive. The explicit benchmarking against industry leaders like GPT-6 Astra and Gemini 4 further emphasizes a competitive push to establish these models as the go-to solution for cutting-edge technical applications.

The introduction of Fable 5.2 and Mythos 5.2, particularly with their enhanced performance and reduced costs, suggests a market trend favoring specialized, high-achieving models for specific, complex use cases over general-purpose assistants. This release challenges the perceived trade-off between capability and cost, indicating that organizations may soon be able to leverage top-tier AI for ambitious, open-ended projects without incurring prohibitive expenses. The availability of a faster, albeit more expensive, mode also caters to different operational needs, highlighting Anthropic's strategy to offer flexibility within their premium offerings.