Anthropic introduced Claude Fable 5.2 and Claude Mythos 5.2 as models for demanding, open-ended technical work, citing improvements in coding, reasoning, vision, and communication. The two share the same underlying model, but Fable includes stronger safeguards against misuse; Mythos is available to organizations working on beneficial cybersecurity or biomedical research through dedicated programs. Anthropic says both models perform as well as or better than recent models on most alignment measures and were evaluated before release by outside groups. They generate outputs up to 30% faster than their predecessors.
The highlighted results span several fields. On the FrontierCode 2.0 benchmark, Fable 5.2 at high effort matched Opus 5.5 at maximum effort for about half the cost per task, while its extra-high setting achieved the top score among released models tested. A thousand Fable agents formalized a computer-checkable proof of Fermat’s Last Theorem in 30 hours, substantially outperforming Opus 5.5 in the same setup. Mythos modeled 28 of 34 attempted gene-therapy capsids, potentially helping biologists design them more quickly and cheaply. Fable also performed strongly at controlling simulated robots, including unfamiliar tasks. Fable 5.2 is available on paid Claude plans, the Claude Platform, and major cloud platforms, priced at $8 per million input tokens and $40 per million output tokens. Anthropic recommends Opus 5.5 for routine coding and knowledge work.
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.