Enhanced reasoning for Agents: Announcing Box AI support for Claude Sonnet 4.6

Anthropic's latest AI model, Claude Sonnet 4.6, showcased significant advancements over its predecessor, Claude Sonnet 4.5, particularly in handling complex, multi-step enterprise workflows, as evaluated in various sectors such as the public sector and healthcare. The model demonstrated notable improvements in reasoning capabilities, achieving a 77% accuracy score, a 15 percentage point increase from Sonnet 4.5, and excelled in orchestrating and performing Q&A over complex tasks. Specific areas of improvement include financial data analysis where Sonnet 4.6 successfully computed investment-to-cost ratios, and integrity in reporting, where it accurately recommended decisions based on correct student data calculations. Additionally, Sonnet 4.6 showed superior performance in extracting structured data from unstructured content, achieving 89% accuracy for fields requiring mathematical logic and over 80% accuracy on various file types like PDFs and Docx files. These enhancements suggest Sonnet 4.6's capacity to automate sophisticated workflows with high precision, making it a robust tool for demanding enterprise applications, now available in Beta in AI Studio and via the Box API.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.