Claude35Sonnet is the latest artificial intelligence model from Anthropic, one of the world's leading AI labs1 The company promises that it will be faster than its predecessor, better understand humor and even be able to read your handwriting

The Claude 3 Opus was already impressive It's the model I called the "most human-like" of AI chatbots I was quick to play with 35Sonnet, but it seems to be more natural and a better understanding of irony Claude is also listed as the best alternative to ChatGPT in my guide to chatbots

This is the first release of the new generation of Claude35 models, followed by updates to tiny Haiku and massive Opus For now, the Sonnet 35 surpasses the Claude 3 Opus in many ratings, despite being faster and cheaper

Anthropic also publishes artifacts on Claude's website This is a new feature that allows Claude to run code snippets and display websites in the sidebar More modalities and memory features are coming soon

The updated model is already available to free and paid users on Claude's website and iOS app Professional subscribers will have a high rate limit as well as ChatGPT

Claude has a 3-tier model similar to Google Gemini The smallest one is Haiku, designed for quick summary tasks, which is very fast Sonnet is the middle tier for most use cases and is available for free Opus is a large model equivalent to GPT-4

In the benchmark test, we surpassed Opus and GPT-4o in graduate level reasoning questions and undergraduate level knowledge

It comfortably outperformed all major models on multilingual mathematics but it outperformed Opus on math problem solving but it was not as good as GPT-4o

One of the biggest upgrades was to its vision features It is the ability to analyze and understand images It could be a graph, chart, photo or handwriting

Anthropic states that text can be accurately transcribed even from incomplete images — which makes captchas redundant

"The Sonnet now surpasses Claude 3 works on all standard vision benchmarks," explained Anthropic "Improvements are most noticeable in tasks that require visual reasoning, such as interpreting charts, graphs, or transcribing text from incomplete images

One of the more surprising updates was the launch of the artifact 1 According to Anthropic, this is a new feature that extends how users interact with Claude 

When a user requests to generate a piece of code, a text document, or a website design, the work is shown off Currently I have to implement it myself to get the code and see if it works, but in the artifact it shows up in the sidebar This also includes the generation of vector graphics and the entire game

It will be displayed in a dynamic workspace area and can be edited and built in real time"AI-generated content is now being previewed in projects and workflows

This new feature, which follows a similar move by OpenAI with Chatgpt, will enable CLAUDE to move from AI chatbot to productivity tools go 

"In the near future, the team, and ultimately the entire organization, will be able to securely centralize knowledge, documentation, and work in progress in one shared space, and Claude will be able to support on-demand team members