A pretraining checkpoint from Anthropic's next-generation Mythos-class model, "numbat", has been leaked, containing a hybrid GDN/MLA architecture with 1,224 experts and NVFP4. The model's architecture and internal consistency match expectations for a Mythos-class model, and a reference implementation is included in the repository. The checkpoint is from a mid-pretraining stage, with 772 shards and a total size of approximately 7.72 TB. AI summary
Firehose
Filtered to Hacker News, tagged “anthropic” · clear filters
Browse: People · Companies · Papers · Podcasts · Hacker News · Deep dives
Browse by tag
Anthropic is expected to receive a high "junk" rating from credit agencies following its initial public offering (IPO), with Blackstone, its anchor investor, predicting a double B range rating, the top tier of speculative "junk" grade. This rating would make it more difficult and costly for Anthropic to raise funds, which has been used to fuel its massive spending on AI development. Anthropic officially filed to go public earlier this week, sparking speculation about its IPO's impact on the broader AI industry. AI summary