Anthropic Launches Claude Fable 5.2 and Mythos 5.2 AI Models With Improved Coding and Reasoning Capabilities

Claude Fable 5.2 was launched by Anthropic on Monday as the Dario Amodei-led AI giant's latest AI model. The Fable series model was introduced along with Mythos 5.2, Anthropic's most advanced cybersecurity-focused AI model. The new Mythos model is available to select cybersecurity professionals via a verification programme. Both models are claimed to bring performance improvements on multiple fronts, including coding, reasoning, and vision. Anthropic says that both new Fable and Mythos models were found to communicate “more clearly” than their predecessors. Moreover, the two AI models are more affordable than Fable 5.1 and Mythos 5.1.

Claude Fable 5.2, Mythos 5.2 Price, Availability

In a blog post, Anthropic has announced that the new Claude Fable 5.2 and Mythos 5.2 AI models are priced at $8 (roughly Rs. 750) per million input tokens and $40 (about Rs. 3,700) per million output tokens. In Fast Mode, the models cost $16 (roughly Rs. 1,500) per million input tokens and $80 (about Rs. 7,500) per million output tokens.

The Claude Fable 5.2 AI model is currently available via Amazon Web Services, Google Cloud, Microsoft Azure, and other platforms. Developers can access the model via the Claude platform. Moreover, for this week, Anthropic is offering higher usage limits to paid users. Meanwhile, Mythos 5.2 can be accessed by select cybersecurity professionals via the Cyber Verification Program, while Enterprise users can get it via Claude Security.

Lastly, Claude Fable 5.2 and Claude Mythos 5.2 with Fast Mode can be accessed by eligible users via the Claude Code, the Claude Platform, Claude Models on Google Cloud, and Microsoft Foundry.

Claude Fable 5.2, Mythos 5.2 Capabilities

The Dario Amodei-led AI lab claims that Claude Fable 5.2 and Mythos 5.2 deliver enhancements on multiple fronts. Both AI models are claimed to generate outputs up to 30 percent faster. The Fast Mode further speeds up token generation by up to 2.5 times. Fable 5.2 outperformed Google's frontier Gemini 4 Carbon AI model on multiple benchmarks.

For reference, on FrontierCode 2.0, testing agentic coding capabilities, Fable 5.2 scored 49.1 percent at the Xhigh effort level, while Gemini 4 Carbon scored 41.1 percent at medium effort. It also outperformed Fable 5.1, Opus 5.5, and OpenAI's GPT-6 Astra, which scored 32.4 percent, 45.4 percent, and 42.8 percent, respectively.

However, Gemini 4 Carbon managed to score higher than Fable 5.2 in handling business workflows on AutomationBench. Coming to Mythos 5.2, the AI model delivered improved performance in detecting cybersecurity vulnerabilities, scoring 89.6 percent on CVE Rediscovery, which is notably higher than Mythos 5.1's score of 78.3 percent.

Further, Anthropic claims that Fable 5.2 can execute “challenging” and complex tasks with fewer tokens than Fable 5.1 on Medium effort. Meanwhile, Mythos 5.2 is claimed to be better at reproducing and fixing vulnerabilities. The company says that a company found that one in six bugs found by Mythos 5.2 were missed by other frontier AI models.