(Bloomberg) -- Anthropic is releasing a new artificial intelligence model that lets users decide if they want a quick answer to a simple question or a more time-consuming response that mimics human reasoning — a novel approach that may help the AI startup stand out in a competitive landscape. With Claude 3.7 Sonnet, users will be able to choose whether to have the AI system spend more or less time computing an answer, depending on the complexity of their query. The model rolled out on Monday to free and paid users, Anthropic said in a blog post, though nonpaying users will initially not be able to use additional computing power to respond to their prompts.In recent months, a growing number of AI startups, including OpenAI, DeepSeek and Elon Musk’s xAI, have introduced new models that can devote more time to computing an answer before responding, a process tech companies typically refer to as reasoning. But while the industry has positioned reasoning systems as the next frontier of AI, Anthropic is betting users may sometimes crave a little more simplicity.“What we’re really trying to do is make it really seamless to adopt this capability where it makes sense, but not have it brought to bear where it doesn’t make sense,” Mike Krieger, Anthropic’s chief product officer, told Bloomberg News.
Most Read from Bloomberg
-
Trump Targets $128 Billion California High-Speed Rail Project
-
Trump Asserts Power Over NYC, Proclaims ‘Long Live the King’
-
NYC’s Congestion Pricing Pulls In $48.6 Million in First Month
-
As Visitors Discover Ghent, the City Is Trying to Prevent a Tourism Takeover
Approaches similar to Anthropic’s may become more common soon. After spending several years releasing more capable AI models at a rapid pace, some artificial intelligence developers are now thinking about ways to make the user experience less complicated.Earlier this month, OpenAI Chief Executive Officer Sam Altman said that his company plans to eventually combine its GPT models, which powered the original ChatGPT chatbot, with its newer o-series of models to build AI systems that can automatically determine how long to ruminate over a query before responding. Eventually, Anthropic may also automate the decision to spend more or less time computing an answer to a query, according to Jared Kaplan, the company’s co-founder and chief science officer.
If a user asks Claude 3.7 Sonnet to spend additional time computing a response, the model will show written details of the chain-of-thought process it follows, a design choice that OpenAI, xAI and DeepSeek have also recently embraced. Kaplan said this method can help users get a better sense for how the model arrives at an answer, and to see where things go wrong when a mistake is made.