Anthropic has launched Claude Fable 5.2 and Claude Mythos 5.2, which the company describes as its most advanced AI models yet. The Anthropic Claude 5.2 launch was announced on November 9, and Anthropic claims the models outperform OpenAI's GPT-6 Astra in several benchmarks while costing 20% less. The company also says the new models generate responses up to 30% faster than their predecessors.
Lower prices and faster responses
Developers who use the API, the interface that lets businesses build their own software on top of an AI model, will pay $8 per million input tokens and $40 per million output tokens. Tokens are the small chunks of text a model reads and writes. The previous models cost $10 and $50 for the same amounts, so the input price falls by $2 per million tokens.
Cached input, which is text the model has already processed and stored for reuse, costs $0.20 per million tokens. GPT-6 Astra charges $1 for the same category, which makes the new rate 80% lower. The launch also includes an optional fast mode that doubles generation speed for double the price. Anthropic has not explicitly linked the price cut to competition with GPT-6 Astra, although the lower prices put it in a stronger position against OpenAI.
Benchmark results
Benchmarks are standard tests used to compare models. Anthropic's published results show Fable 5.2 ahead of Opus 5.5, another model in the comparison, on nearly every metric the company reported. On FrontierCode 2.0, which tests agentic coding, meaning AI carrying out multi-step programming tasks on its own, Fable 5.2 scored 49.1%. Opus 5.5 scored 45.5% and GPT-6 Astra scored 42.8%.
On the RoboDojo robotics benchmark, Fable 5.2 reached 64.9%, more than double the 27.6% recorded for Fable 5.1. These figures come from the company itself, and no independent testing was reported at the time of the announcement. Anthropic also says the models improve coding, reasoning, scientific research and image understanding, and those gains are likewise its own claims.
Two versions, one underlying technology
Both models share the same underlying technology but differ in safety restrictions. Fable 5.2 includes added cybersecurity and biology safeguards, which allows wider availability to general users. Mythos 5.2 lifts those restrictions for vetted organisations working in cybersecurity and life sciences.
A summary post about the launch listed availability through paid plans, Claude Code, the API, and AWS, Google Cloud and Azure. The prices announced are in US dollars, and no rupee prices were given.
Aasan alfaaz mein
Anthropic ne naye AI models Claude Fable 5.2 aur Mythos 5.2 launch kiye hain, jo 9 November ko announce hue. Company ke mutabiq yeh pehle se tez hain aur inki API qeemat pehle se kam hai. Fable 5.2 aam users ke liye hai, jabkay Mythos 5.2 sirf munzoor shuda organisations ke liye hai. Aam logon ke liye iska seedha asar abhi wazeh nahi, lekin AI tools banane walon ki lagat kam ho sakti hai.
