On Monday, Anthropic announced the release of updated versions of two of its most powerful artificial intelligence (AI) models. The announcement was made ahead of an initial public offering (IPO) - the company's stock market listing, which is expected in the week of November 16.
The new models are named Fable 5.2 and Mythos 5.2. According to the company, they run up to 30% faster than previous versions, especially when performing complex tasks such as programming, while costing 20% less. Both models are built on a single base architecture.
The difference between them lies in the level of restrictions. Fable 5.2 is equipped with safeguards that block most cybersecurity and biology tasks, so the company offers it to all users. Mythos 5.2 has fewer such restrictions, and access to it is open only to vetted organizations working in cybersecurity or biomedical research.
The first versions of Fable and Mythos were released in June amid controversy. A few days after launch, both models were withdrawn: the US Department of Commerce, citing national security considerations, required Anthropic to block access for foreign nationals. The restrictions were lifted several weeks later.
The new release coincided with preparations for the IPO, which is expected to raise up to 100 billion dollars and become the largest in history. Anthropic is headquartered in San Francisco. The company's market value after listing is estimated to exceed 1 trillion dollars, although in recent weeks forecasts have declined from a peak of 2 trillion.
The offering is seen as an indicator of how long the AI investment boom supporting the US economy will last. The company is showing rapid growth: third-quarter revenue reached 17 billion dollars, compared with 1.4 billion a year earlier, according to documents filed with the Securities and Exchange Commission earlier this month.
However, sales growth has slowed compared with previous quarters due to tough competition from OpenAI (the creator of ChatGPT) and cheaper models, many of which are made in China. In addition, in the third quarter Anthropic recorded an operating loss of 567 million dollars, linked to increased spending on short-term computing capacity.
The company's valuation was also affected by a general market downturn after a series of quarterly reports that worried investors. These reports raised questions about the hundreds of billions of dollars that US tech giants are spending on AI. An additional factor was concerns about possible tighter regulation of the tech industry, including data center construction, after Democrats regained control of Congress in the US midterm elections. Measures restricting data center construction could slow the growth plans not only of Anthropic but also of its competitors - OpenAI, Meta, Microsoft and Google.