Anthropic, a leader in AI safety and research, has officially announced the launch of Claude, the first model in its next-generation Claude 3.5 family. Released on June 20, 2024, this new model sets a new standard for intelligence, speed, and cost-effectiveness in the rapidly evolving landscape of artificial intelligence.
Claude 3.5 Sonnet demonstrates remarkable improvements across a wide array of benchmarks. Notably, it outperforms Anthropic's previous flagship model, Claude 3 Opus, in many evaluations, including graduate-level reasoning (GPQA), undergraduate-level knowledge (MMLU), and coding proficiency (HumanEval). This is particularly impressive given that Claude 3.5 Sonnet operates at twice the speed of Opus, making it highly suitable for complex, enterprise-scale AI applications and latency-sensitive tasks. The model is also more affordable, priced at $3 per million input tokens and $15 per million output tokens, coupled with a generous 200,000 token context window.
Beyond text-based tasks, the new model showcases significantly enhanced capabilities in vision understanding. It has established new industry benchmarks for visual reasoning, outperforming Claude 3 Opus on standard vision tests. This allows Claude 3.5 Sonnet to accurately interpret charts and graphs, and precisely transcribe text even from imperfect or distorted images, a critical capability for industries like retail, logistics, and financial services where visual data is abundant.
A groundbreaking addition accompanying this release is the new Artifacts feature on the Claude.ai web interface. This innovative tool allows users to interact with Claude's generations in a more dynamic way. When a user requests content such as code snippets, text documents, or even website designs, these creations appear in a dedicated window alongside the chat. Users can then see, edit, iterate upon, and seamlessly integrate these artifacts into their projects and workflows, effectively transforming Claude.ai into a collaborative workspace.
Anthropic has emphasized that Claude 3.5 Sonnet maintains the company's strong commitment to AI safety and privacy. The model has undergone rigorous testing to mitigate misuse and ensure responsible deployment.
Accessibility is broad: Claude 3.5 Sonnet is available for free on Claude.ai and the Claude iOS app, with higher rate limits for Claude Pro and Team plan subscribers. Developers can also access the model through the Anthropic API, Amazon Bedrock, and Google Cloud's Vertex AI platform. This release signifies a major step forward, offering users and developers a powerful, efficient, and versatile AI tool. Anthropic plans to release Claude 3.5 Haiku and Claude 3.5 Opus later this year, further expanding the capabilities of the Claude family.