Anthropic released Claude 3.5 Sonnet and Artifacts, a new way to collaborate with Claude

Claude 3.5 Sonnet is the first release in the Claude 3.5 family, delivering intelligence and performance surpassing even Claude 3 Opus and other competitor models in several standard evaluations while retaining the speed and cost of its predecessor, Claude 3 Sonnet. Claude 3.5 Sonnet establishes new industry benchmarks for GPQA, MMLU, and HumanEval, which test graduate-level reasoning, undergraduate-level knowledge, and coding proficiency.

According to Anthropic, Claude 3.5 Sonnet is better than Claude 3 Sonnet at understanding nuance, humor, complex instructions, and writing high-quality content. Internal testing shows that Claude 3.5 Sonnet has remarkable coding capabilities, including fixing a bug in or adding a feature to a codebase, independently performing some coding tasks, and handling code translations. It is also a strong vision model, surpassing Claude 3 Opus' performance and showing proficiency at tasks requiring chart and graph interpretation. Additionally, Claude 3.5 Sonnet can transcribe text from images, a key capability to deliver insights in several of its use cases.

Claude 3.5 Sonnet is now available free at Claude.ai and the iOS app. It can also be accessed via the Anthropic API, Amazon Bedrock, and Google Cloud’s Vertex AI. Claude 3.5 Sonnet is priced at $3 per million input tokens and $15 per million output tokens, with a 200K token context window. Additionally, the web experience at Claude.ai has been enriched with Artifacts, a feature that enables a dedicated window to organize the content users have asked Claude to generate, such as code snippets, text documents, or website designs. This results in a dynamic workplace that allows users to edit and build upon Claude's generations easily and efficiently.

Anthropic has subjected Claude 3.5 Sonnet to rigorous testing and external expert evaluations. The company plans to release Claude 3.5 Haiku and Claude 3.5 Opus later this year, promising continuous improvements in AI intelligence, speed, and cost-effectiveness.