In-Short
- Anthropic launches Claude 3.5 Sonnet, surpassing its own Claude 3 Opus in performance.
- Available on Claude.ai, iOS app, and through various APIs, priced at $3 per million input tokens.
- Enhanced capabilities in graduate-level reasoning, knowledge, coding, and vision tasks.
- Introduces Artifacts feature for collaborative AI interaction, with a strong focus on safety and privacy.
Summary of Claude 3.5 Sonnet Launch
Anthropic has introduced its latest AI model, Claude 3.5 Sonnet, which outshines its predecessor, Claude 3 Opus, and competitors in various benchmarks. This mid-tier model is now freely accessible on Claude.ai, the Claude iOS app, and through the Anthropic API, Amazon Bedrock, and Google Cloud’s Vertex AI. With a pricing model based on input and output tokens, Claude 3.5 Sonnet offers a 200K token context window and operates at double the speed of its previous version.
The new model boasts superior performance in graduate and undergraduate-level reasoning and knowledge tasks, coding proficiency, and vision capabilities. It is particularly adept at understanding nuances, humor, and complex instructions, and excels in producing high-quality, natural-toned content. Claude 3.5 Sonnet’s vision skills are evident in tasks that require visual reasoning, such as interpreting charts and graphs, and it can transcribe text from imperfect images, which is beneficial for various industries.
Anthropic has also rolled out a new feature called Artifacts on Claude.ai, which allows users to interact more collaboratively with the AI by viewing, editing, and building upon generated content in real-time. Despite these advancements, the company maintains a strong commitment to safety and privacy, with rigorous testing and training protocols to reduce misuse. External experts, including the UK’s AI Safety Institute and Thorn, have contributed to refining the model’s safety mechanisms.
Anthropic reassures users about privacy, clarifying that generative models are not trained on user-submitted data without explicit permission, and to date, no customer data has been used for training purposes.
Image Credit: Anthropic
Explore Further
For more detailed insights into Anthropic’s Claude 3.5 Sonnet and its capabilities, visit the original source.
Footnotes
Image Credit: Anthropic
