SAN FRANCISCO — The race to build the smartest artificial intelligence chatbot completed another lap on Monday when startup Anthropic unveiled a new AI model it says is more advanced than those before it.
San Francisco-based Anthropic, an OpenAI rival started by former leaders at the ChatGPT maker, announced Claude 3 — the third generation of its large language model.
While OpenAI has closely tied itself to its business partner Microsoft, Anthropic’s primary cloud computing partner is Amazon, which has said it would invest up to $4 billion in the company. Another cloud provider, Google, has also invested in Anthropic.
All of the partnerships have attracted scrutiny from U.S. antitrust enforcers at the Federal Trade Commission, which is expecting more information later this month in response to compulsory orders it sent the companies.
The new Claude 3 models come in three versions, all of which have computer vision capabilities that enable them to analyze what’s in a photo, chart or graph. But they won’t generate new images, avoiding the troubles that forced Google to recently shut down a feature of its Gemini chatbot over how it was depicting race and ethnicity.
The most advanced and priciest, called Claude 3 Opus, beats OpenAI’s ChatGPT-4 and Google’s Gemini Ultra in certain multiple-choice tests that are used to measure a chatbot’s capabilities, according to Anthropic.
