High-profile AI startup Anthropic on Monday released a new version of its Claude chatbot, announcing it outperformed other leading chatbots in various standard benchmark tests, including systems from Google and OpenAI. did.

Anthropic CEO and co-founder Dario Amodei said the new technology, called Claude 3 Opus, will be particularly useful for analyzing scientific data and generating computer code.

Anthropic is one of the few companies at the forefront of generative AI, the technology that instantly creates text, images, and audio. Dr. Amodei and his other Anthropic founders helped pioneer this technology while working as researchers at OpenAI, a startup that launched the generative AI boom with the release of the chatbot ChatGPT in late 2022.

Chatbots like ChatGPT can answer questions, write term papers, and generate small computer programs. Just like humans, they can also generate false or misleading information.

When OpenAI released a new version of its technology called GPT-4 last spring, it was widely considered to be the most powerful chatbot technology in use by both consumers and businesses. Google recently introduced a comparable technology called Gemini.

But major artificial intelligence companies are distracted by one controversy after another. There is a shortage of computer chips needed to build AI. And it faces countless lawsuits over how it collects digital data, another essential element in creating AI (The New York Times sued Microsoft and OpenAI over use of copyrighted material) ).

Still, technology continues to advance at an astonishing pace.

Anthropic claims its Claude 3 Opus technology outperforms both GPT-4 and Gemini in areas such as mathematical problem solving, computer coding, and general knowledge.

Claude 3 Opus will be available starting Monday to consumers who pay $20 a month for a subscription. A less powerful version called Claude 3 Sonnet is available for free.

The company enables businesses to build their own chatbots and other services using Opus and Sonnet technology.

Both versions of the technology can respond to images as well as text. These can, for example, analyze flowcharts or solve mathematical problems involving diagrams and graphs.

However, this technique cannot generate images. Google recently stopped its ability to generate human faces after Gemini created images of people of color wearing German World War II uniforms.



Source link