Anthropic has unveiled Claude Fable 5.2 and Claude Mythos 5.2, its latest large language models (LLMs), which feature significant upgrades in coding, reasoning, and performance. The announcement was made today, with Fable 5.2 being more widely available and equipped with enhanced cybersecurity and biology safeguards, while Mythos 5.2 is restricted to verified organizations for sensitive tasks.
Fable 5.2 and Mythos 5.2 Performance Enhancements
Fable 5.2 has shown impressive results, outperforming its predecessor, Fable 5.1, by over 10 percent across 14 artificial intelligence benchmarks. Notably, it achieved a score of 74.6 percent on the Terminal-Bench-Science 0.1 test, marking a 22 percent improvement. The model’s capabilities were demonstrated when it reduced the time to generate a computer-verifiable version of a well-known mathematical proof from 11 days to just 30 hours by utilizing 1,000 agents for parallel calculations.
In a separate evaluation, Mythos 5.2 successfully modeled 28 out of 34 protective shells of adeno-associated viruses (AAVs), showcasing its potential for biological research. This achievement is significant as it surpasses the capabilities of other advanced models like Google DeepMind’s AlphaFold2, which can only model individual proteins.
Pricing and Accessibility of New Models
Both Fable 5.2 and Mythos 5.2 are priced at USD 8 per million input tokens and USD 40 per million output tokens, which is a 20 percent reduction compared to previous versions. An optional high-performance mode is available, offering responses 2.5 times faster than the standard tier for double the cost. Anthropic aims for these models to assist in complex technical tasks, while recommending its earlier model, Opus 5.5, for routine coding and knowledge tasks.
The launch of these models comes as Anthropic prepares for a potential initial public offering (IPO) later this month. The company’s strategy includes limiting access to Mythos 5.2 to verified organizations, particularly for handling sensitive cybersecurity issues.
Background on Anthropic’s Innovations
Founded by former OpenAI researchers, Anthropic focuses on developing AI technologies that prioritize safety and usability. The company has been at the forefront of creating advanced language models that cater to both general and specialized applications, reflecting the growing demand for AI solutions across various industries.
These developments in AI technology are expected to attract more users and enhance productivity in technical fields, especially as organizations seek to automate complex processes more efficiently.
