Claude Sonnet 5.5 launched with faster performance and lower costs for professional tasks

Claude Sonnet 5.5, an artificial intelligence model developed by Anthropic and released on September 28, is designed to perform professional tasks more quickly and at a lower operational cost. The update includes enhancements in programming, information analysis, and the creation of documents, presentations, and spreadsheets.

According to the company, the model generates responses over 30% faster than its predecessor, Sonnet 5, and can reduce costs by up to 30% per task. Results vary depending on the activity and the complexity of the request.

Creating spreadsheets and presentations

One of Claude Sonnet 5.5’s key applications is transforming information into structured materials for professional use. The model can help organize data in spreadsheets, prepare presentations, and create documents with more sophisticated visual layouts.

For example, a professional can provide market research data and request a spreadsheet organized by categories, along with a presentation highlighting key findings. They can also specify a visual template to guide slide creation.

While the tool can save time in preparing these materials, it’s essential to verify numbers, sources, and conclusions before use.

Document and information analysis

The model has also been enhanced to handle knowledge-based tasks that require reading, comparing, and interpreting information.

In practice, professionals can use it to summarize lengthy reports, compare documents, identify data trends, and structure information to support decision-making.

Anthropic reports that Sonnet 5.5 has shown improvements in professional task evaluations across various occupations. However, the company’s published results do not guarantee the same performance in all scenarios.

Programming and application development

Programming is another area where the new version demonstrates significant progress. Claude Sonnet 5.5 can assist in identifying code errors, implementing functionalities, and executing multi-step development tasks.

This capability also enables technically skilled users to develop prototypes of digital tools, such as interactive pages, small applications, and data visualization dashboards.

Instead of requesting a specific code snippet, users can describe what they want to build, test the result, and request adjustments. However, applications intended for real-world use must undergo testing and technical review.

Interpreting images and graphics

The model also includes improvements in understanding visual information, such as interpreting graphs, recognizing patterns in images, and using visual references to perform specific tasks.

For instance, an analyst can submit a performance graph and request an explanation of key variations. A marketing professional can ask for an analysis of the visual organization of a presentation or webpage.

The accuracy of interpretation depends on the quality of the provided material. Poorly legible images, inappropriate scales, and incomplete data can compromise the results.

Executing multi-step tasks

Another feature of Sonnet 5.5 is its ability to handle tasks that require a sequence of actions, rather than responding to a single question.

This type of functionality is associated with so-called agentic AI, which uses tools and executes steps to achieve a goal defined by the user.

For example, a user could request the analysis of a dataset, the identification of patterns, and the creation of a report with the results. Depending on the environment and available tools, the model can perform parts of this workflow in an integrated manner.

Executing longer tasks does not eliminate the need for supervision. It’s important to verify the information used, monitor the actions performed, and ensure the result aligns with the initial goal.

How to develop skills for using AI at work

Understanding a tool’s functions is only part of the learning process. To effectively apply artificial intelligence in professional settings, it’s also necessary to know how to formulate instructions, evaluate responses, and identify tasks where the technology can enhance productivity.

How much does it cost to use Claude Sonnet 5.5?

According to Anthropic, the model is available on Claude and can also be accessed by developers through programming platforms and cloud services.

For the API, which is used to integrate artificial intelligence into systems and applications, the pricing is $2 per million input tokens and $10 per million output tokens. Tokens are units of text processed by the model, and the final cost depends on the volume used.

The company also offers temporary context storage features, known as prompt caching, which can reduce costs in certain workflows.