Anthropic has delivered a resounding response in the frontier artificial intelligence (AI) race, with the Claude Fable 5.2 and Mythos 5.2 models. These come soon after Google’s Gemini 4 Argon models set the frontier benchmark. The launches provide a timely boost to Anthropic’s momentum ahead of its expected market debut later this month. Fable, for instance, can coordinate more than 1,000 AI agents simultaneously.
Anthropic has mounted a strong challenge in the frontier artificial intelligence (AI) race with the launch of its Claude Fable 5.2 and Mythos 5.2 models, coming soon after Google’s Gemini 4 Argon models set a new benchmark in the field. The launches provide a timely boost to Anthropic’s momentum ahead of its expected market debut later this month. Fable, for instance, can coordinate more than 1,000 AI agents simultaneously.
Anthropic says Fable 5.2 exhibits improvements in areas like coding, reasoning, and vision, whereas Mythos 5.2 is their most capable model yet for cybersecurity engineering. “Testers have found that Fable 5.2 and Mythos 5.2 communicate more clearly than past Fable and Mythos models. They generate outputs up to 30% faster than Claude Fable 5.1 and Claude Mythos 5.1, and they’re available at a lower price: $8 per million input tokens and $40 per million output tokens,” Anthropic says, in a statement.
What can Fable 5.2, Mythos 5.2 do
Fable 5.2 and Mythos 5.2 are best suited for coding and knowledge tasks. Benchmark scores peg Fable 5.2 not only as a significant generational step forward over the Fable 5.1, but also Gemini 4 Carbon and OpenAI’s GPT6 Astra.
In the agentic coding FrontierSWE v2 benchmark, Fable 5.2 scored 69.9% while Fable 5.1 had logged 56.3% in the same tests. It is also ahead of GPT-6 Astra (65.5%) and Gemini 4 Carbon (61.2%). In the AutomationBench business workflow benchmark, Gemini 4 Carbon leads with a 53.9% score, with Fable 5.2 logging a close second at 51.4%, and ahead of GPT-6 Astra (41.4%).
In the scientific agentic research Terminal-Bench-Science 0.1 benchmark, Mythos 5.2 sets the new goalpost with a score of 74.6%, ahead of GPT-6 Astra (68.1%) and Gemini 4 Carbon (64.8%).
Anthropic says that as an experiment, one Anthropic engineer gave Fable 5.2 the task of designing a prototype operating system. Starting from an empty folder, the model coordinated over 1,000 subagents, working around the clock for eight days writing, testing, and reviewing code. It produced a working operating system that boots on a laptop, with its own core that runs the machine, drivers for the disk, keyboard, and network, permanent file storage, windows, a login screen, an installer, and built-in Claude functionality.
Anthropic says Fable 5.2 and Mythos 5.2 performed as well as, or better than their recent models on most alignment measures.The model was also tested before release by external evaluators, including the AI Security Institute, Frontier Design and METR.
In another example, Anthropic says gene therapies can be delivered in a protein shell called a capsid. Mythos 5.2 found a new way to predict a whole capsid’s structure. “In our tests, it accurately modeled 28 of the 34 capsids it attempted. Biologists could use it to design capsids faster and more cheaply,” they say.
Fable 5.2 also shows substantially improved skills in controlling physical robots, even on tasks it has never explicitly been trained to do.On RoboDojo, a benchmark with a simulated two-arm robot, it outperformed other models and most AI systems built to control robots-score of 64.9% compared with GPT-6 Astra’s 29.8% and Fable 5.1’s 27.6%.
The Fable 5.2 model is now available on all paid Claude subscription plans. They say that Fable and Mythos are the same underlying model. Fable has strong safeguards to guard against misuse. Organisations who intend to develop applications in cybersecurity or biomedical research can apply for Mythos through the Life Sciences and Cyber Verification Programs.