SuperAPI AI Directory
Partner with us中文AI newsFavorites
DirectoryNews and tutorials

Claude Sonnet 4.5 - Anthropic's latest AI coding model

Author: AI小集 · Original source: AI工具集

What is Claude Sonnet 4.5

Claude Sonnet 4.5 is Anthropic's latest and most powerfulcodingmodel. The model performs excellently in multiple fields such as coding, computer operation, reasoning, and mathematics, topping the SWE-bench Verified test and able to focus on work for over 30 hours. The model has strong agentic capabilities and can automate complex tasks through APIs and tools. Claude Sonnet 4.5 has significantly improved alignment and safety, reducing undesirable behavior and enhancing defense capabilities. Claude Sonnet 4.5 supports a variety of application scenarios, including code generation, knowledge Q&A, and content creation, etc. Currently, Claude Sonnet 4.5 is fully available, including claude.ai the official website,App and API (API pricing is the same as the previous generationClaude Sonnet 4).

Claude Sonnet 4.5 introduces the Imagine with Claude temporary research preview feature for users. Through natural language instructions, users can interact with Claude and generate and adjust software code and functionality in real time. All content is created instantly without the need for preset code. This feature is currently only available to Max subscribers, aiming to accelerate software development, prototyping, and creative exploration, and provide an efficient and flexible interactive experience.

Claude Sonnet 4.5

Claude Sonnet 4.5's main features

  • Long-running agents: Provides excellent instruction following, tool selection, error correction, and advanced reasoning capabilities, suitable for customer-facing agents and complex AI workflows.
  • Code generation: Sonnet 4.5 can complete tasks across the entire software development lifecycle, from initial planning to bug fixing, maintenance, and large-scale refactoring, supporting up to 64K output tokens, suitable for complex code generation and planning.
  • Browser and computer operation: The model leads in computer operation capabilities and can reliably handle various browser-based tasks, from competitive analysis to procurement processes to customer onboarding. Its computer operation capabilities are more accurate and reliable than the previous generation.
  • Cybersecurity: Teams using Sonnet 4.5 and Claude Code can deploy agents to automatically patch vulnerabilities to prevent exploitation, shifting from passive detection to proactive defense.
  • Financial analysis: The model can handle all tasks from basic financial analysis to advanced predictive analytics, such as continuously monitoring global regulatory changes and adjusting compliance systems in advance, realizing the transformation from manual audit preparation to intelligent risk management.
  • Business tasks: The model performs excellently in creating and editing office files such as slides, documents, and spreadsheets.
  • Research: Sonnet 4.5 can search external and internal data sources to synthesize comprehensive insights in complex information environments.
  • Content generation and analysis: The model performs excellently in writing, can understand nuances and tone, generate more engaging content, and analyze content at a deeper level.

Claude Sonnet 4.5's core upgrades

  • Claude Code: Claude Sonnet 4.5 adds checkpoint functionality in Claude Code, able to save progress and instantly roll back to previous states, while also updating the terminal interface and releasing a native VS Code extension to meet strong user demand and improve the development experience.
  • Claude API: Claude Sonnet 4.5 adds context editing functionality and memory tools to the Claude API, enabling agents to run for longer and handle more complex tasks, further enhancing the API's functionality and application scenarios.
  • Claude app: Claude Sonnet 4.5 integrates code execution and file creation capabilities (including spreadsheets, slides, and documents) directly into conversations in the Claude app, providing a Chrome extension for Max users who joined the waitlist last month, enhancing convenience and efficiency in real-world use.
  • Claude Agent SDK: Claude Sonnet 4.5 launches the Claude Agent SDK, providing developers with the infrastructure and tools used to build Claude Code, enabling developers to build their own frontier products.
Claude Sonnet 4.5

Claude Sonnet 4.5's Performance

  • SWE-bench Verified Evaluation: Claude Sonnet 4.5 reaches industry-leading performance in the SWE-bench Verified evaluation. In practical applications, the model can stay focused on complex, multi-step tasks for over 30 hours.
  • OSWorld Benchmark Test: In the OSWorld benchmark test, Claude Sonnet 4.5's performance improves significantly, reaching 61.4% accuracy, ahead of other models.
  • Chrome Browser: Through the Claude for Chrome extension, Sonnet 4.5 can work directly in the browser, including browsing websites, filling out spreadsheets, and completing tasks.
  • Broad Evaluations in Reasoning and Mathematics: Claude Sonnet 4.5 demonstrates improved capabilities in broad evaluations such as reasoning and mathematics, performing better in these areas compared to previous models.
  • Domain-Specific Knowledge and Reasoning Ability: Experts in finance, law, medicine, and STEM find that Sonnet 4.5 shows significant improvements in domain-specific knowledge and reasoning ability compared to older models (including Opus 4.1).
Claude Sonnet 4.5

Claude Sonnet 4.5's Application Scenarios

  • Long-Running Agents: Suitable for complex tasks that need to run for a long time, such as automated workflows, continuous monitoring, and multi-step task processing.
  • Code Generation: Throughout the entire software development lifecycle, from planning to implementation, maintenance, and refactoring, supporting complex programming tasks.
  • Browser and Computer Operation: Supports handling browser-based tasks, such as automated web operations, data scraping, and online process handling.
  • Cybersecurity: Deployed agents can automatically detect and fix vulnerabilities, shifting from passive defense to proactive security protection.
  • Financial Analysis: The model can perform financial analysis from basic to advanced, including risk assessment, compliance monitoring, and predictive analysis.

Claude Sonnet 4.5's Official Report

CakeGrowth QR Code

Free ClaimClaude Sonnet 4.5 Official Model Report, scan the WeChat QR code to follow and reply:20250930

Claude Sonnet 4.5's Target Users

  • Software Developers: The model's powerful code generation and optimization capabilities help developers quickly implement complex programming tasks and improve development efficiency.
  • Cybersecurity Teams: The model can automatically detect and fix vulnerabilities, helping teams shift from passive defense to proactive security protection.
  • Financial Analysts: The model can handle financial analysis from basic to advanced, including risk assessment and compliance monitoring, supporting intelligent risk management.
  • Enterprise Business Personnel: Efficiently create and edit office files, such as slides, documents, and spreadsheets, improving daily office efficiency.
  • Researchers: Integrate internal and external data sources to generate comprehensive insights across complex information domains, supporting academic and business research.