Claude Sonnet 3.5 debuts with cutting-edge vision and code execution tools

· 4 min read
Artifacts on Claude

Anthropic has released Claude Sonnet 3.5, which outperforms OpenAI's GPT-4o. This new model is available for free on Claude and includes additional vision features. It can now execute code and has improved vision capabilities. The Artifact feature is the first publicly available preview.

React contact form in Claude's playground
React contact form in Claude's playground

Claude 3.5 generated React JSX code for a simple contact form and successfully ran it in the Artifacts playground. Users can switch between Preview and Code modes in the playground. The Preview mode allows running almost any code as it functions as a fully-featured React playground. Claude also enables users to switch between different code revisions when modifications are requested.

Rate limits on free accounts for Claude 3.5 are strict, with quotas running out in about 30 minutes of use. Despite this, users can create projects, such as a simple snake game, to revisit later. The upcoming Starting feature will allow users to treat their chats as projects, making it more important for managing ongoing work.

Claude is available on multiple platforms, including the web and iOS. Users are considering creating multiple accounts to manage their usage better.

Claude 3.5 Sonnet Launch:

Anthropic has introduced Claude 3.5 Sonnet, the first model in their upcoming Claude 3.5 family. This new model sets a new industry standard for AI intelligence, outperforming both competitor models and its predecessor, Claude 3 Opus, across various evaluations.Key Features and Improvements:

  1. Intelligence: Claude 3.5 Sonnet excels in graduate-level reasoning (GPQA), undergraduate-level knowledge (MMLU), and coding proficiency (HumanEval).
  2. Speed: It operates at twice the speed of Claude 3 Opus, making it ideal for complex tasks like context-sensitive customer support and multi-step workflows.
  3. Natural Language Processing: The model shows significant improvement in understanding nuance, humor, and complex instructions.
  4. Content Creation: Claude 3.5 Sonnet is exceptional at producing high-quality content with a natural, relatable tone.
  5. Coding Capabilities: In an internal agentic coding evaluation, Claude 3.5 Sonnet solved 64% of problems, compared to Claude 3 Opus's 38%. It can independently write, edit, and execute code with advanced reasoning and troubleshooting abilities.
  6. Vision Capabilities: This model surpasses Claude 3 Opus on standard vision benchmarks, excelling in visual reasoning tasks like interpreting charts and graphs.
Introducing Claude 3.5 Sonnet
Introducing Claude 3.5 Sonnet—our most intelligent model yet. Sonnet now outperforms competitor models and Claude 3 Opus on key evaluations, at twice the speed.

Anthropic Blog

Availability and Pricing:

  • Claude 3.5 Sonnet is available for free on Claude.ai and the Claude iOS app.
  • Claude Pro and Team plan subscribers have access with higher rate limits.
  • It's also accessible via the Anthropic API, Amazon Bedrock, and Google Cloud's Vertex AI.
  • Pricing: $3 per million input tokens and $15 per million output tokens.
  • Context Window: 200K tokens.
Claude Sonnet 3.5 performance
Claude Sonnet 3.5 performance (source: Anthropic)

New Feature - Artifacts:

Anthropic has introduced "Artifacts" on Claude.ai, a feature that allows users to see, edit, and build upon Claude's creations in real-time. This transforms Claude from a conversational AI into a collaborative work environment.

Safety and Privacy:

  • Claude 3.5 Sonnet has undergone rigorous testing and training to reduce misuse.
  • It maintains an ASL-2 safety level, as determined by red teaming assessments.
  • Anthropic has engaged with external experts, including the UK's Artificial Intelligence Safety Institute (UK AISI), for pre-deployment safety evaluation.
  • The company emphasizes its commitment to privacy, stating that they do not train generative models on user-submitted data without explicit permission.

Future Developments:

Anthropic plans to release Claude 3.5 Haiku and Claude 3.5 Opus later this year to complete the Claude 3.5 model family. They are also working on new modalities, features for businesses, and exploring capabilities like Memory for more personalized user experiences.

LaTeX on Claude
LaTeX on Claude

Besides Artifacts, Claude is working on another Feature Preview which will be quite useful for Math-related tasks: LaTeX support. However, this feature is not specific to Claude 3.5 Sonnet and has been in development for a while. But it is already working as expected!

LaTeX on Claude
LaTeX on Claude

About Claude: Claude is an AI model developed by Anthropic. It is designed to assist with various tasks, including code generation and execution. The model is available for free and includes advanced vision features.