Anthropic Unveils Claude 3.5 Sonnet
A new GPT-4o competitor has arrived!
Created on June 20|Last edited on June 20
Comment
Anthropic has launched Claude 3.5 Sonnet, the first model in the new Claude 3.5 family, available for free on Claude.ai and the Claude iOS app. Subscribers of the Claude Pro and Team plans can access it with higher rate limits, and it is also available through the Anthropic API, Amazon Bedrock, and Google Cloud’s Vertex AI. Claude 3.5 Sonnet, priced at $3 per million input tokens and $15 per million output tokens, features a 200K token context window.
Performance and Capabilities
Claude 3.5 Sonnet excels in graduate-level reasoning, undergraduate-level knowledge, and coding proficiency, showing significant improvements over previous models like Claude 3 Opus. It operates at twice the speed of Claude 3 Opus, making it ideal for complex tasks such as context-sensitive customer support and orchestrating multi-step workflows.
In internal evaluations, Claude 3.5 Sonnet solved 64% of coding problems, a marked improvement over Claude 3 Opus's 38%. It can independently write, edit, and execute code, handling code translations with ease, making it effective for updating legacy applications and migrating codebases.
Vision Model and New Features
Claude 3.5 Sonnet also advances in vision capabilities, surpassing Claude 3 Opus in tasks requiring visual reasoning, like interpreting charts and graphs, and accurately transcribing text from imperfect images, benefiting retail, logistics, and financial services sectors.
Additionally, Anthropic introduced Artifacts on Claude.ai, allowing users to generate and edit content such as code snippets, text documents, or website designs in a dynamic workspace. This feature marks Claude’s evolution into a collaborative work environment, with future plans to support team collaboration.
Long Context Performance
Claude 3.5 Sonnet consistently demonstrates high recall rates, maintaining strong performance even as the context length grows. It outperforms other models, particularly noticeable at longer context lengths, where its recall remains near 1.00.
Commitment to Safety and Privacy
Anthropic emphasizes safety and privacy, with Claude 3.5 Sonnet undergoing rigorous testing to mitigate misuse. The model remains at ASL-2, and Anthropic has engaged external experts, including the UK Artificial Intelligence Safety Institute, to refine safety mechanisms. The company ensures that no user-submitted data is used for training without explicit permission.
Future Developments
Anthropic plans to release Claude 3.5 Haiku and Claude 3.5 Opus later this year, aiming to improve the balance between intelligence, speed, and cost. They are also exploring features like Memory for personalized user experiences and developing new modalities and enterprise application integrations.
Claude 3.5 Sonnet represents a significant step forward in AI capabilities, with ongoing improvements and new features on the horizon to enhance user experience and application scope. Users are encouraged to provide feedback to help shape future developments.
Add a comment
Tags: ML News
Iterate on AI agents and models faster. Try Weights & Biases today.