Anthropic, a startup with backing from tech giants Google and Amazon.com, has introduced a new suite of artificial intelligence models named Claude 3, intensifying the competition among Silicon Valley firms to develop increasingly powerful AI technology.
According to Anthropic, the flagship model in the Claude 3 family, Claude 3 Opus, surpasses competitors such as OpenAI’s GPT-4 and Google’s Gemini 1.0 Ultra in various benchmark tests, making it the top performer in its class, Reuters news report said.
“This is the Rolls-Royce of models, at least at this point in time,” stated CEO Dario Amodei in an interview, highlighting the superior capabilities of Claude 3 Opus.
While advancements from rival companies have been announced or anticipated, Reuters could not independently verify Anthropic’s claims. The release of Claude 3 follows the launch of Anthropic’s Claude 2 in July, underscoring the ongoing race among tech companies to dominate AI performance rankings, even as businesses grapple with integrating such advanced technology into their operations.
Daniela Amodei, President of Anthropic, emphasized that customers would be willing to pay a premium for Claude 3 Opus if they required AI capabilities for complex tasks, such as intricate financial analysis.
Anthropic disclosed pricing details for Claude 3 Opus, stating that it would charge $15 for processing every 1 million tokens of data, with smaller models available at significantly lower rates. In comparison, OpenAI’s GPT-4 Turbo model charges $10 per million tokens.
In a notable development, Anthropic highlighted Claude 3 as its first “multimodal” AI suite, capable of processing both text and image inputs. CEO Amodei explained that while Claude 3 can analyze images, it does not generate them due to lower demand from enterprise clients.
This distinction comes in the wake of Google’s recent suspension of image-generation features for its Gemini chatbot, following concerns about inaccuracies and issues related to diversity.
Anthropic announced plans to make Claude 3 models accessible through Amazon and Google’s cloud platforms, as well as directly selling access in 159 countries. This move reflects the company’s commitment to expanding the reach of its advanced AI technology to a global audience.
