Although it’s not as popular as its competitor ChatGPT, the Claude AI model, which is said to be just as capable by its users, has exciting developments on the way. With the new generation of Claude, the daily tasks you perform on your computer will now be easily handled by the AI model. Here are the details and latest updates…
Claude 3.5 Sonnet Beta is coming too!
With the latest announcement from Anthropic, the creator of Claude, it seems like a new era is about to begin in the world of AI. While significantly improving the current Claude 3.5 Sonnet model, the company is also preparing to launch a brand new model called Claude 3.5 Haiku to users. But the real bombshell was the announcement of the beta release of the computer usage feature.
Anthropic’s new beta feature allows Claude to interact with your computer just like a human. Claude, which can see the screen, move the mouse, and use the keyboard, will be able to automate daily tasks such as filling out forms, browsing websites, and gathering data.
This new feature scored 14.9 percent in OSWorld tests, performing more than twice as successfully as other AI systems. When more time was given to the AI model to complete tasks, this rate rose to 22 percent.
The updated Claude 3.5 Sonnet has also made significant progress in the field of software engineering. With a 49 percent success rate in SWE-bench Verified tests, it managed to surpass its competitors. It also achieved impressive results in TAU-bench evaluations, with scores of 69.2 percent in the retail sector and 46 percent in the airline industry.
The new 3.5 Haiku is designed with a focus on balancing speed and cost. Haiku has outperformed even the previous generation large model Claude 3 Opus in some tests, especially showing remarkable performance in coding. With a 40.6 percent success rate in SWE-bench Verified tests, it has surpassed models like the original Claude 3.5 Sonnet and GPT-4o.
Haiku will be available on Amazon Bedrock and Google Cloud’s Vertex AI platforms later this month. Initially, the model will serve as text-only, but it will gain the ability to generate visuals in the future.
What do you think about this? Have you had the chance to try Claude before? Don’t forget to share your thoughts in the comments.
{{user}} {{datetime}}
{{text}}