Amazon Bedrock is the best place to build and scale generative ai applications with large language models (LLM) and other foundation models (FM). It enables customers to leverage a variety of high-performance FMs, such as Anthropic's Claude family of models, to create custom generative ai applications. Looking ahead to 2021, when Anthropic began building on AWS, no one could have imagined how transformative the Claude family of models would be. We've been making next-generation generative ai models accessible and usable for businesses of all sizes through Amazon Bedrock. In just a few months since Amazon Bedrock became generally available on September 28, 2023, more than 10,000 customers have been using it for deliveries, and many of them are using Claude.. Clients like ADP, Broadridge, Cloudera, Dana-Farber Cancer Institute, Genesys, Genomics England, GoDaddy, Intuit, M1 Finance, Perplexity ai, Proto Hologram, Rocket Companies and more are using Anthropic's Claude models on Amazon Bedrock to drive innovation in generative ai and create transformative customer experiences. And today we announce an exciting milestone with the arrival of the next generation of Claude to Amazon Bedrock: Claude 3 Opus, Claude 3 Sonnet and Claude 3 Haiku.
Introducing the Claude 3 models from Anthropic
Anthropic introduces its next generation of Claude with three advanced models optimized for different use cases. Haiku is the fastest and most profitable model on the market. It is a fast compact model for almost instant responsiveness. For the vast majority of workloads, Sonnet is 2x faster than Claude 2 and Claude 2.1 at higher intelligence levels. It excels at intelligent tasks that require quick responses, such as knowledge retrieval or sales automation. And it strikes the ideal balance between intelligence and speed, qualities especially critical for enterprise use cases. Opus is the most advanced, capable and next-generation FM with deep reasoning, advanced mathematics and coding capabilities, with top-notch performance on highly complex tasks. You can navigate open prompts and novel scenarios with remarkable fluidity, including automating tasks, generating hypotheses, and analyzing charts, graphs, and forecasts. And Sonnet is available for the first time on Amazon Bedrock today. Anthropic's current evaluations suggest that the Claude 3 family of models outperforms comparable models on the Mathematical Word Problem Solving (MATH) and Multilingual Mathematics (MGSM) benchmarks, critical benchmarks used today for the LLMs.
- Vision capabilities – Claude 3 models have been trained to understand structured and unstructured data in different formats, not only language, but also images, charts, diagrams and more. This allows companies to create generative ai applications that integrate diverse multimedia sources and solve truly cross-domain problems. For example, pharmaceutical companies can view drug research articles along with protein structure diagrams to accelerate discovery. Media organizations can automatically generate image captions or video scripts.
- Best-in-class benchmarks – Claude 3 outperforms existing models in standardized assessments such as math problems, programming exercises, and scientific reasoning. Customers can optimize domain-specific experimental procedures in manufacturing or audit financial reports based on contextual data, automated and with high accuracy using ai-powered responses.
Specifically, Opus outperforms its peers on most common evaluation benchmarks for ai systems, including undergraduate level expert knowledge (MMLU), graduate level expert reasoning (GPQA), basic mathematics ( GSM8K) and more. Demonstrates high levels of understanding and fluency in complex tasks, leading the frontier of general intelligence.
- reduced hallucination – Companies require predictive and controllable results from artificial intelligence systems that drive automated processes or customer interactions. Claude 3 models mitigate hallucinations using constitutional ai techniques that provide transparency into the model's reasoning, as well as improving accuracy. Claude 3 Opus shows an estimated twofold gain in accuracy over Claude 2.1 on difficult open-ended questions, reducing the likelihood of wrong answers. Since Claude is trusted by enterprise clients in industries such as healthcare, finance, and legal research, reducing hallucinations is essential for safety and performance. The Claude 3 family sets a new standard for reliable generative ai production.
Benefits of the Anthropic Claude 3 FM on Amazon Bedrock
Through Amazon Bedrock, customers will have easy access to build with Anthropic's newest models. This includes not only natural language models but also its expanded range of multimodal ai models capable of performing advanced reasoning on text, images, graphics, and more. Our collaboration has already helped customers accelerate the adoption of generative ai and delivered business value. Below are some ways customers have been using Anthropic's Claude models on Amazon Bedrock:
“We are developing a generative ai solution on AWS to help customers plan epic trips and create life-changing experiences with personalized travel itineraries. Developing with Claude at Amazon Bedrock, we reduced itinerary generation costs by nearly 80% when we quickly built a secure, scalable ai platform that can organize our book content in minutes to provide consistent, highly accurate travel recommendations . “We can now repackage and personalize our content in multiple ways across our digital platforms, based on customer preferences, while highlighting trusted local voices, just as Lonely Planet has done for 50 years.”
— Chris Whyde, Senior Vice President of Data Science and Engineering, Lonely Planet
“We are working with AWS and Anthropic to host our custom, fine-tuned Anthropic Claude model on Amazon Bedrock to support our strategy to rapidly deliver generative ai solutions at scale and with cutting-edge encryption, data privacy, and secure ai technology built into everything.” what do we do. “Our new Lexis+ ai platform technology features conversational search, detailed summaries and intelligent legal writing capabilities, enabling lawyers to increase their efficiency, effectiveness and productivity.”
— Jeff Reihl, Executive Vice President and Chief technology Officer, LexisNexis Legal & Professional
“At Broadridge, we have been working to automate understanding of regulatory reporting requirements to create greater transparency and increase efficiency for our clients operating in domestic and global financial markets. By using Claude on Amazon Bedrock, we are excited to get even greater accuracy in our experiments with processing and summarization capabilities. With Amazon Bedrock, we have choice in the use of LLM and value the performance and integration capabilities it offers.”
— Saumin Patel, VP of Generative ai Engineering, Broadridge
The Claude 3 family of models meets diverse needs, allowing customers to choose the model that best suits their specific use case, which is key to developing successful prototype and subsequent production systems that can deliver real impact. , whether for a new product, feature, or process that improves the end result. With customer needs in mind, Anthropic and AWS deliver services where they matter most for organizations of all sizes:
- Improved performance – Claude 3 models are significantly faster for real-time interactions thanks to optimizations in hardware and software.
- Greater precision and reliability – Thanks to massive scaling and new self-monitoring techniques, the expected double gains in accuracy for complex questions in long contexts mean that ai is even more useful, safe and honest.
- Easier and safer customization – Customization capabilities, such as Retrieval Augmented Generation (RAG), simplify training models on proprietary data and create applications supported by diverse data sources, so customers get ai tailored to their unique needs. Additionally, proprietary data is never exposed to the public Internet, never leaves the AWS network, is transferred securely over VPC, and is encrypted in transit and at rest.
And AWS and Anthropic continually reaffirm our commitment to advancing generative ai responsibly. By constantly improving the capabilities of the model by engaging with frameworks such as Constitutional ai or the ai” target=”_blank” rel=”noopener”>White House Voluntary Commitments on aiwe can accelerate the safe and ethical development and deployment of this transformative technology.
The future of generative ai
Looking ahead, customers will create entirely new categories of ai-powered generative applications and experiences with the latest generation of models. We've only just begun to harness the potential of generative ai to automate complex processes, augment the human experience, and reshape digital experiences. We expect to see unprecedented levels of innovation as customers choose Anthropic models augmented with multimodal capabilities that leverage all the tools they need to build and scale generative ai applications on Amazon Bedrock. Imagine sophisticated conversational assistants that provide fast, highly contextual responses, personalized recommendation engines that seamlessly combine relevant images, diagrams and associated insights to intuitively guide decisions. Imagine scientific research driven by a generative ai capable of reading experiments, synthesizing hypotheses, and even proposing novel areas for exploration. There are many possibilities that can be realized by taking full advantage of everything that generative ai has to offer through Amazon Bedrock. Our collaboration ensures that companies and innovators around the world have the tools to reach the next frontier of ai-powered generative innovation responsibly and for the benefit of all.
Conclusion
It's still early days for generative ai, but strong collaboration and a focus on innovation are ushering in a new era of generative ai on AWS. We can't wait to see what customers build next.
Resources
Please see the following resources for more information about this announcement:
About the Author
Swami Sivasubramanian He is vice president of data and machine learning at AWS. In this role, Swami oversees all AWS database, analytics, artificial intelligence, and machine learning services. His team's mission is to help organizations put their data to work with a complete end-to-end data solution to store, access, analyze, visualize and predict.