Shopping cart

Hype Pakistan is your digital pulse on the stories shaping our nation and the global stage. We deliver the "hype" that matters, keeping you ahead of the curve on breakthroughs in Tech, Startups, Business, and Global shifts. By bridging the gap between local momentum and international innovation, we provide the curated insights you need to turn information into opportunity.

Anthropic Launches Its Most Powerful Model At 20% Cheaper Price

Claude Fable 5.2

Anthropic has released Claude Fable 5.2 and Claude Mythos 5.2, its most powerful AI models, at a price about 20% lower than the versions they replace. The company announced both models on Monday, November 9, 2026.

For developers, the new models cost $8 for every million words-pieces of input (called tokens) and $40 per million tokens of output. The older Fable 5.1 and Mythos 5.1 cost $10 and $50.

In rupees, at roughly Rs. 277 to the dollar, that is about Rs. 2,200 per million input tokens and Rs. 11,100 per million output tokens. A million tokens is roughly 750,000 English words.

One model, two versions

Fable 5.2 and Mythos 5.2 are the same underlying AI. The difference is in the safety rules.

Fable 5.2 has extra safeguards around cybersecurity and biology. It is available on paid Claude plans, in Claude Code, on Anthropic’s developer platform, and through Amazon Web Services, Google Cloud and Microsoft Azure.

Mythos 5.2 lifts those safeguards. It is only open to organisations approved through Anthropic’s cybersecurity and life sciences verification programmes, such as security researchers and drug-discovery labs.

Faster, with fewer refusals

Anthropic says the new models produce answers up to 30% faster than the 5.1 versions and explain things more clearly.

An optional “fast mode” speeds up replies by up to 2.5 times, but costs double: $16 for input and $80 for output per million tokens. Developers who can wait for results can use batch processing at half the normal price.

Anthropic also said Fable 5.2 blocked about 55% fewer requests than Fable 5.1 in early testing. A common complaint about the earlier version was that it refused harmless questions that looked like security or biology topics.

Claims still being tested

Anthropic says Fable 5.2 matched the top score of its cheaper Opus 5.5 model on an expert-level knowledge test called Humanity’s Last Exam, at less than one-tenth of the cost per task.

It also said about 1,000 copies of Fable 5.2 worked together for 30 hours to turn the proof of Fermat’s Last Theorem, a famous maths problem, into code a computer can check. Outside mathematicians have not yet reviewed that work, Technology.org noted.

These results come from Anthropic’s own testing. Independent checks usually follow in the weeks after a launch.

A busy race

The launch comes about two months after the 5.1 versions arrived on September 1, 2026, and a month after Anthropic released its smaller Haiku 5.5 model. Reuters reported that Anthropic is expanding its range ahead of a planned stock market listing, as OpenAI, Google and Meta push rival models.

Per token, Fable 5.2 is still twice the price of Opus 5.5, which costs $4 and $20. Anthropic recommends starting with Opus 5.5 and moving to Fable only for harder work.

Leave a Reply

Your email address will not be published. Required fields are marked *

Related Posts