Captured photo
Claude 3.5 Sonnet
30

About

Claude 3.5 Sonnet is a versatile generative AI tuned for creative writing, advanced coding, visual reasoning, and autonomous multi-step workflows. Built for enterprises and developers and available via platforms such as Yale’s Clarity, Amazon Bedrock, Vertex AI, and Claude.ai, Sonnet helps teams prototype software, migrate legacy code, draft polished content, and extract insights from images and mixed data. Its improved software engineering performance (SWE-bench Verified rising from 33% to 49%) and benchmark gains on domain tests make it one of the leading public models for practical coding tasks. Sonnet can write, edit, and execute code to accelerate development cycles, support debugging, and generate working functions from high-level specifications. Its vision capabilities interpret charts, diagrams, and imperfect images reliably, enabling better business intelligence in retail, logistics, and finance. Unique tool-integration training enables Sonnet to interact with software interfaces and perform web navigation and multi-step workflows autonomously where platform integrations allow. On Claude.ai, the Artifacts workspace supports collaborative creation and iterative editing of code, documents, and other assets in real time. The model also excels at complex customer support, orchestrating multi-stage resolutions, and producing statistical visualizations and actionable analysis from unstructured data. Operationally, Sonnet delivers lower latency and stronger instruction following than prior versions while maintaining the same price and speed. It underwent safety evaluations by U.S. and U.K. AI safety institutes and follows responsible scaling practices. Limitations include restricted file-upload support on some platforms, non-recommendation for processing ePHI, and the need for human oversight for mission-critical or highly specialized tasks. For teams seeking a high-performance assistant that bridges creative, analytical, and engineering workflows, Claude 3.5 Sonnet offers a powerful, practical, and collaborative solution.

Percs

Fast generation
Multi-modal
High quality
High accuracy

Settings

Diversity control (Top P)-  Controls response diversity by filtering based on cumulative probability. Lower values (0.1-0.5) produce focused, deterministic responses. Higher values (0.7-1.0) allow more creative and diverse outputs.
Temperature-  Controls randomness and creativity. Lower values (0.1-0.5) make responses more focused and deterministic. Higher values (0.6-1.0) increase creativity and variability. Use low for code, high for creative writing.
Response length-  Maximum number of tokens (words/pieces) in the AI's response. Higher values allow longer, more detailed answers but cost more. 1000 tokens ≈ 750 words.
Context length-  Maximum input length the model can process in tokens. Higher values let you include more conversation history or longer documents but use more resources. Max: 200,000 tokens.