Dario Amodei, co-founder and chief executive officer of Anthropic, at the World Economic Forum in 2025.
Stefan Wermuth | Bloomberg | Getty Images
Anthropic on Monday announced its latest artificial intelligence model: Claude Sonnet 4.5.
The model is better at coding, using computers and meeting practical business needs, and it excels in specialized fields like cybersecurity, finance and research, Anthropic said. The Amazon-backed startup, which is valued at $183 billion, is making Claude Sonnet 4.5 available to all users.
Anthropic said Claude Sonnet 4.5 is the “best coding model in the world” according to industry benchmarks like SWE-bench Verified, a test set that measures an AI system’s software coding abilities.
“People are just noticing with this model, because it’s just smarter and more of a colleague, that it’s kind of fun to work with it when encountering problems and fixing them,” Jared Kaplan, Anthropic’s co-founder and chief science officer, told CNBC in an interview.
The model generates higher-quality code, is better at identifying code improvements and can follow instructions more reliably, the company said.
Claude Sonnet 4.5 comes after Anthropic launched Claude Opus 4.1 in August and Claude Sonnet 4 in May. It’s the latest example of the breakneck pace of innovation within the AI industry.
Anthropic was founded by a group of former OpenAI researchers in 2021, and the two companies have been fierce competitors ever since.
OpenAI kick-started the generative artificial intelligence boom following the release of its ChatGPT chatbot in 2022. The startup, which has seen its valuation swell to $500 billion, announced its latest model, GPT-5, in August. The rollout was rocky, as some users complained about losing access to the company’s prior models.
Mike Krieger, Anthropic’s chief product officer, said that Claude Sonnet 4.5 will be the default for users and that Anthropic recommends the model for “basically every use case.”
Even so, users will have options. Paid subscribers can still choose to use Opus, and users with specific workflows can select an older generation of Sonnet if they aren’t ready to migrate overnight, he said.
Claude Sonnet 4.5 is smaller than Claude Opus 4.1 but smarter than it in “almost every single way,” Krieger added.
Anthropic’s Claude Sonnet 4.5 chatbox.
Courtesy of Anthropic
“We have found it, and our customers are finding it, very useful for real, actual work,” Krieger said.
Claude Sonnet 4.5 can run autonomously for 30 hours, and Anthropic said it’s able to maintain focus on complex, multistep tasks throughout that period. Claude Opus 4, which the company launched in May, could run autonomously for just seven hours.
Anthropic has also been able to improve the model’s behavior through extensive safety training, the company said. It has reduced “concerning behaviors” like deception, power seeking and sycophancy, which is when a model tells a user what they want to hear.
Claude Sonnet 4.5 is also more resistant to prompt injection attacks, where a model can get tricked into doing something malicious, like exposing sensitive data.
“This is the biggest jump in safety that I think we’ve seen in the last probably year, year and a half,” Kaplan said.
Additional model launches are already on the horizon for Anthropic. Kaplan said better models are coming, including “very likely Opus.”
“No promises,” he said. “But I think that we’ll probably have one or two more releases before the end of the year.”
WATCH: Microsoft to buy AI from Anthropic in shift away from OpenAI, sources say
