At Oracle, we’re focused on creating products that help enterprises quickly and easily embed state-of-the-art AI innovations into their business processes to get the most value out of their data and drive high-value business outcomes. Oracle is uniquely positioned to make this proposition a reality for our customers through a comprehensive AI strategy that integrates our strengths in enterprise-grade data management with new approaches that bring AI to data.
We’ve deeply embedded AI and machine learning (ML) across our entire product suite, spanning databases, AI infrastructure, cloud services, applications, and industry solutions. We’ve continued our work around maintaining data security and governance. And we’ve built an AI infrastructure used by the builders of the world’s biggest and most capable large language models (LLMs).
Today, we’re announcing expanded capabilities for Oracle AI that further simplify the development and deployment of enterprise-grade generative AI solutions. These enhancements include the availability of the Cohere Command R family of models on the Oracle Cloud Infrastructure (OCI) Generative AI service, new features to improve the quality and usability of generative language models, new models for OCI Data Science AI Quick Actions, and expanded availability of our portfolio of platform services for AI including text-to-speech for OCI Speech.
New generative AI models and features: Innovating for enterprise adoption
Cohere Command R and Command R+ models
We’re pleased to announce the general availability of Cohere Command R on OCI Generative AI for on-demand and dedicated hosting. Command R is a highly scalable and secure enterprise-grade AI model. It offers industry-leading multilingual capabilities and leverages retrieval-augmented generation (RAG) to excel at business use cases with high accuracy and verifiability.
Additionally, we’re offering limited availability of Cohere Command R+, which provides even stronger accuracy for customers who need the highest quality generative outputs. Cohere recently introduced this family of scalable models that strike a balance between high-quality output, strong accuracy, and efficient performance for enabling enterprises to move from proof of concept into production-grade AI. Learn more about the new Command R offering on OCI Generative AI.
“We’re excited to have Cohere’s latest enterprise-grade Command R model series available on OCI for enterprises to deploy at production scale,” said Vinod Devan, head of global partnerships at Cohere. “Our Command R model series is state-of-the-art in business-critical capabilities, such as advanced retrieval-augmented generation, multilingual support for global business operations, and tool use to automate business workflows. We look forward to the opportunity to integrate our new models across Oracle to boost productivity for customers.”
Tabnine, the AI code assistant that accelerates and simplifies software development while keeping your code private, secure, and compliant, is a new OCI customer using OCI Generative AI. By selecting the Cohere Command R model as an underlying LLM that powers Tabnine Chat, Tabnine is providing its users with capabilities such as creating tests and documentation; and explaining, fixing, generating, and maintaining code.
“By using the Cohere Command R model running on Oracle Cloud Infrastructure, Tabnine delivers our users with another option for best-in-class AI infrastructure,” said Peter Guagenti, president of Tabnine. “As a trusted resource for millions of developers, we are committed to maintaining the highest standards of privacy, security, and performance. The Cohere Command R model on OCI Generative AI is a welcomed addition and we are excited to see it in action.”
Meta Llama3-70B
We now also support the pretrained Meta Llama 3 70B model. This model offers reasoning, code generation, and instruction capabilities that are greatly improved from the previous Llama 2 model. With this release, you can access the Llama 3 70B model through on-demand or dedicated hosting. You can fine-tune the Llama 3 70B model with your own custom data set to enhance the precision of the model with the low-rank adaptation (LoRA) fine-tuning method.
LlamaIndex
OCI Generative AI is now integrated with LlamaIndex, the leading framework for building LLM applications. Creating chatbots, agents, and other applications with generative AI is now simpler.
New Chat API
We introduced the new Chat API, which supports the Command R and Llama 3 models, accessible via the OCI SDK, API, or console Playground. The new Chat API provides an easy integration into the new models by matching the format of the partner model APIs. It also supports Tool Use for function calling or building complex agents.
OCI Generative AI Playground
We’ve made improvements to our user interface with a new conversational experience in the OCI Generative AI Playground. This new interface makes it easier for users to interact with the models and explore their capabilities.
New limited availability for OCI Generative AI Agents
In January, we announced beta availability for OCI Generative AI Agents, a new service that combines the power of LLMs and RAG with your enterprise data. Today, we’re pleased to announce limited availability for the service. OCI Generative AI Agents removes barriers to using enterprise data strategically, retrieves real-time data on the fly, and provides information in natural language. Learn more about the new agent offering on OCI Generative AI Agents.
New features for OCI Data Science AI Quick Actions
OCI Data Science AI Quick Actions is generally available, providing users with a no-code option to access a range of open source LLMs, including ones from Meta and Mistral AI. This feature enables users to deploy AI models more quickly and easily. The OCI Data Science service provides a suite of tools and resources for data scientists and ML engineers to develop and deploy AI models.
AI Quick Actions now supports importing any model from OCI Object Storage. In addition, the OCI Data Science team has tested and verified the following models:
- Meta-Llama-3-8b/8b-instruct/70b/70b-instruct
- gemma-1.1-7b-it
- gemma-2b/2b-it/7b
- codegemma-1.1-2b/7b-it
- codegemma-2b/7b
- Falcon-40b-instruct
- ELYZA-japanese-Llama-2: 7b/7b-instruct/13b/13b-instruct
Learn more about the new AI Quick Actions offerings in OCI Data Science.
OCI Speech: Real-time transcription and text to speech
Neural text to speech in OCI Speech is in limited availability. This feature enables the generation of human-like speech from text, which can be used for various applications, such as creating content, improving accessibility, and translating languages.
Recently, we announced the limited availability of real-time transcription in OCI Speech, a new feature that enables users to convert audio streams to text in real time. Real-time transcription is critical for many use cases, including healthcare, call centers, and media. This capability will be especially useful for enhancing the efficiency of medical professionals who require accurate real-time transcription of patient orders. Learn more about the new text-to-speech feature on OCI Speech.
OCI Language: Healthcare domain models
We’re augmenting our OCI Language service with new models that are now in general availability for the healthcare domain. These models are designed to help process and analyze clinical documents, such as electronic health records, progress notes, and clinical trial notes. The models include a Protected Health Information option, which can identify and deidentify private health information in text documents, as well as named entity recognition for recognizing clinical entities, such as symptoms. Learn more about the new models on OCI Language.
Conclusion
The latest advancements in our Oracle AI portfolio demonstrate our continued commitment to delivering best-in-class AI services and infrastructure to our customers.
AI has the potential to transform the enterprise. However, to be truly useful in solving business problems, it must be enterprise-grade, secure, accurate, and accessible. Oracle’s latest AI offerings are designed to meet these needs. They’re powered by high performance and cost-effective infrastructure and seamlessly embedded across our industry-leading cloud applications and suite of data management services. With Oracle, enterprises can harness the power of generative AI to improve customer experiences, boost employee productivity, and drive innovation.
We can’t wait to see what our customers create with these new capabilities.
To learn more, visit:
To get started, visit:
The preceding is intended to outline our general product direction. It is intended for information purposes only, and may not be incorporated into any contract. It is not a commitment to deliver any material, code, or functionality, and should not be relied upon in making purchasing decisions. The development, release, timing, and pricing of any features or functionality described for Oracle’s products may change and remains at the sole discretion of Oracle Corporation.