Claude Fable 5.2 is presented as a model for demanding reasoning and extended agentic tasks, including hours-long agent sessions, multistep research, and work that ends in a polished document, spreadsheet, or presentation. It supports a one-million-token context window and outputs of up to 128,000 tokens, with adaptive thinking always enabled and high effort as the default. Its listed latency is moderate. Input and output cost $8 and $40 per million tokens, respectively; prompt-cache writes cost $10 per million tokens for five-minute storage or $16 for one-hour storage, while cache reads cost $0.20. Batch API requests receive a 50% discount on input and output.
The model accepts text and images and has a June 2026 knowledge and training-data cutoff. Compared with Opus 5.5, it costs twice as much at the listed rates and defaults to higher effort; Sonnet 5.5 is faster and cheaper, while Haiku 5.5 is the lowest-cost option. Mythos 5.2 has the same specifications and pricing but is restricted to organizations approved through Anthropic verification programs. Fable 5.2 launched November 9, 2026, and is active, with retirement promised no earlier than November 9, 2027. Practical limitations include mandatory 30-day data retention unless Anthropic expressly authorizes otherwise, errors for forced tool-use settings, and errors when sampling parameters such as temperature are supplied. Users control thinking depth through the effort parameter.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.
Researchers tested 16 LLMs on all 55 TIS-100 puzzles, giving each one attempt and time to refine its code. The models found 133 new performance tradeoffs across 18 puzzles and set three outright cycle records, though human players improved on most solutions within days. (tis-100.wincon.dev)
Australia is considering a new criminal offence for AI companies that fail to take reasonable steps to prevent harm caused by their models. The government is also weighing an AI safety watchdog with power to impose significant fines, following a hack of a Medicare statistics portal involving OpenAI. (abc.net.au)
AI is increasingly becoming a gateway through which people find information. Audiences still resist having AI produce their news. (niemanlab.org)
Nvidia holds a central position in the AI industry, making disruption to the company potentially catastrophic for the wider ecosystem. The race for advanced AI is also intensifying economic and geopolitical conflict, while concentrated corporate power and infrastructure face growing political resistance. (france24.com)
LLM inference can require reading about 26 GB of model weights to generate each token from a 13-billion-parameter model. The explanation shows how model size drives memory bandwidth demands during text generation. (youtube.com)
Deletion Tuning matches full fine-tuning across three LLM post-training methods by zeroing fewer than 0.05% of model weights. Its compact masks also preserve prior-task accuracy better than full fine-tuning, LoRA, and random sparse tuning. (arxiv.org)
Today's best Hacker News stories, summarized and screenshotted, one email a day.