Claude logo
Boonyawee Sirimaya
min read
June 24, 2024

Claude 3.5 Sonnet: Redefining AI Performance and Efficiency

Anthropic has unveiled Claude 3.5 Sonnet, the first member of its new Claude 3.5 model family, which is poised to redefine the benchmarks of artificial intelligence. This advanced model not only surpasses competitor models but also outperforms its predecessor, Claude 3 Opus, in a variety of evaluations. Claude 3.5 Sonnet uniquely combines high-level intelligence with remarkable speed and cost-efficiency, setting a new standard in the AI industry.

Performance comparison table of AI models Claude 3.5 Sonnet, Claude 3 Opus, GPT-4o, Gemini 1.5 Pro, and Llama-400b across various benchmarks including reasoning, knowledge, coding, math, and problem-solving tasks.
Performance comparison of AI models across various cognitive and computational tasks

Broad Accessibility and Competitive Pricing

Claude 3.5 Sonnet is available for free on platforms such as and the Claude iOS app. Those subscribed to the Claude Pro and Team plans can benefit from significantly higher rate limits. The model can also be accessed via the Anthropic API, Amazon Bedrock, and Google Cloud’s Vertex AI. With a pricing structure of $3 per million input tokens and $15 per million output tokens, and a generous 200K token context window, Claude 3.5 Sonnet offers cutting-edge AI capabilities at a competitive rate.

Outstanding Performance Metrics

Claude 3.5 Sonnet sets new industry benchmarks in areas such as graduate-level reasoning (GPQA), undergraduate-level knowledge (MMLU), and coding proficiency (HumanEval). The model demonstrates a superior understanding of nuances, humor, and complex instructions, producing high-quality, relatable content. Operating at twice the speed of Claude 3 Opus, Claude 3.5 Sonnet is well-suited for demanding applications, including sophisticated customer support and complex workflow management.

Enhanced Coding Abilities

In a recent internal evaluation, Claude 3.5 Sonnet excelled by solving 64% of coding problems, a significant improvement over the 38% success rate of Claude 3 Opus. This evaluation tested the model’s ability to debug or enhance open-source codebases based on natural language descriptions. Equipped with advanced tools and instructions, Claude 3.5 Sonnet can independently write, edit, and execute code, demonstrating sophisticated reasoning and problem-solving skills. The model is also adept at code translations, making it invaluable for modernizing legacy applications and migrating codebases.

Superior Visual Processing

Claude 3.5 Sonnet also excels in visual reasoning, outperforming Claude 3 Opus on standard vision benchmarks. This includes tasks such as interpreting charts and graphs, as well as accurately transcribing text from imperfect images. These capabilities are particularly beneficial for industries like retail, logistics, and financial services, where AI can derive more insights from visual data compared to text alone.

Introducing Artifacts: A Revolutionary Collaboration Tool

Artifacts, a new feature on, revolutionizes user interaction with AI. When users request content such as code snippets, text documents, or website designs, these Artifacts appear in a dedicated window alongside the conversation. This creates a dynamic workspace where users can see, edit, and build upon AI-generated content in real-time, integrating it seamlessly into their projects and workflows.

This feature marks a significant evolution from a simple conversational AI to a comprehensive collaborative tool. is set to expand further, supporting team collaboration and enabling organizations to centralize their knowledge, documents, and ongoing work in a shared space, with Claude acting as an on-demand assistant.

Emphasis on Safety and Privacy

Anthropic places a high priority on the safety and privacy of its AI models. Claude 3.5 Sonnet has undergone rigorous testing to minimize misuse and remains at ASL-2 according to red teaming assessments. Collaboration with external experts has enhanced the safety mechanisms. Recently, the UK’s Artificial Intelligence Safety Institute (UK AISI) evaluated Claude 3.5 Sonnet, with findings shared with the US AI Safety Institute (US AISI) under a Memorandum of Understanding between the two organizations.

Input from child safety experts at Thorn and other external advisors has been integrated to refine models and classifiers, ensuring robust safeguards against potential misuse. Furthermore, privacy is a core principle, with a strict policy of not using user-submitted data to train models unless explicitly permitted by users. To date, no customer or user-submitted data has been used in training generative models.

Future Developments

Anthropic is committed to continually enhancing the balance between intelligence, speed, and cost. Later this year, the Claude 3.5 model family will be completed with the releases of Claude 3.5 Haiku and Claude 3.5 Opus. Alongside these developments, new features and modalities are being explored to expand AI use cases for businesses, including integrations with enterprise applications. One notable feature in development is Memory, which will enable Claude to remember user preferences and interaction history, providing a more personalized and efficient AI experience.

User feedback is highly encouraged to shape the ongoing development of Claude. Insights can be submitted directly within the product, helping to inform the development roadmap and enhance the user experience. The innovative ways users employ Claude to build, create, and discover new possibilities are eagerly anticipated.

Consult with our experts at Amity Solutions for additional information on Amity Bots here