Claude Opus is Anthropic’s most intelligent new model. The company says it outperformed rivals such as OpenAI’s GPT-4 in several tasks.
Just a year after Anthropic first made its AI chatbot Claude available to businesses as an alternative to OpenAI’s ChatGPT, it has once again issued a major challenge to its big-name rival.
Anthropic is an AI start-up with backing from Google’s parent company Alphabet. It was founded by former OpenAI employees who have been working on a suite of tools to take a chunk of OpenAI’s market share.
Today (4 March), Anthropic debuted a trio of tools called Claude 3 Opus, Haiku and Sonnet. It claimed that the suite sets new industry benchmarks across a wide range of cognitive tasks.
Opus and Sonnet are now available in 159 countries, while Haiku will be “available soon”, according to a blogpost that accompanied today’s announcement.
Claude 3 Opus, which the company says is its most intelligent model, reportedly outperformed rival models on basic mathematics tasks and reasoning tasks.
Its potential uses include task automation, research and development, and strategy forecasting. Anthropic claims it can navigate open-ended prompts with “remarkable fluency” and “human-like understanding”.
Claude 3 Sonnet, meanwhile, is intended for enterprise teams. It can perform functions such as data processing, quality control, forecasting and product recommendations.
When it becomes available, Claude 3 Haiku will answer queries quickly and catch out potential content moderation flaws. It is billed as a responsive chatbot that can answer questions in a way that mimics humans.
The trio of tools were tested with Asana and Airtable. They are faster and more intelligent than previous versions of Claude chatbots. Haiku, Sonnet and Opus are Anthropic’s first multimodal tools, meaning users can get answers from the unstructured data they upload.
Find out how emerging tech trends are transforming tomorrow with our new podcast, Future Human: The Series. Listen now on Spotify, on Apple or wherever you get your podcasts.
