(Paris) The American company Anthropic announced Thursday the launch of an improved version of its generative artificial intelligence model, Claude, just a few months after the release of its latest series.
Smarter and “twice as fast”, this is the promise of Claude 3.5 Sonnet, the new version of its mid-size model, according to a company press release.
Anthropic presented Claude 3 in March, which includes three gradual power models, Haiku, Sonnet and Opus.
This new version of Sonnet “has the capacity to perform tasks of very high complexity,” Michael Gerstenhaber, product manager at Anthropic, explained to AFP.
“It is particularly powerful for tasks related to coding, for those related to improving intelligent customer service, for supporting data scientists, for data visualization and the creation of literary content,” he said. -he adds.
As with previous models, users will have free access to it on claude.ai, the web version of the AI assistant and on the mobile application (on Apple devices). For businesses, the “Claude Team” plan is paid with secure access to the model.
Claude 3.5 Sonnet will be accessible in the United States, but also in Europe, particularly in France.
The success of OpenAI’s ChatGPT since the end of 2022 has launched the generative AI revolution, which allows machines to understand and produce text, images, sound, lines of code, etc., upon simple request in everyday language.
Founded by former OpenAI employees, Anthropic is trying to distinguish itself from its competitors by putting in place stricter safeguards for this technology that excites and worries.
“Creating systems that are not only efficient, but also reliable, secure and aligned with human values is a complex challenge,” recognized Dario Amodei, boss of Anthropic, quoted in the press release.
This approach had made Claude less impressive than ChatGPT. But the start-up now believes it has managed to get around the obstacle without sacrificing its values.