Anthropic rolls out faster, cheaper AI models ahead of US$100b Wall Street debut

San Francisco-based AI company Anthropic announced on November 10 the release of updated versions of two of its most powerful artificial intelligence models. The new models, Fable 5.2 and Mythos 5.2, are faster and cheaper than previous iterations. Fable 5.2, which has safeguards blocking most cybersecurity and biology tasks, can perform tasks up to 30% faster and costs 20% less than its predecessor.

Mythos 5.2, with fewer restrictions, is available only to vetted organizations in cybersecurity or biomedical research. Both models have been met with controversy since their June launch. The US Commerce Department initially barred foreign nationals from accessing the models due to national security concerns, but this was lifted weeks later.

Anthropic aims to hold its highly anticipated IPO the week of November 16, potentially raising up to US$100 billion, which would make it the largest IPO on record. The company's valuation is expected to exceed US$1 trillion. Anthropic's growth has been significant, with revenue reaching US$17 billion in the third quarter, up from US$1.4 billion a year earlier.

However, sales growth has slowed amid competition from OpenAI and cheaper models, many from China. The company also posted an operating loss of US$567 million in the third quarter due to higher costs for computing capacity. The IPO is seen as a barometer for the longevity of the AI spending boom lifting the US economy, but concerns about investor sentiment and potential regulatory crackdowns from Democrats controlling Congress may impact Anthropic's growth prospects.