Generative AI

Hosted applications in OCI Generative AI now support persistent file storage. Store model artifacts, intermediate results, checkpoints, and ...

You can now select No authentication when creating a hosted application in OCI Generative AI. Use this ...

Sandboxes are now available in OCI Generative AI. A sandbox provides a managed, temporary Linux workspace where ...

OCI Generative AI now supports access to on-demand models through routing profiles in 30 OC1 commercial regions. Call ...

You can now import the Qwen/Qwen3.6-35B-A3B-FP8 model into OCI Generative AI, create endpoints for it, and use ...

You can now import the nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-NVFP4 model into OCI Generative AI, create endpoints for it, and use ...

You can now import the Qwen/Qwen3.8-Flash-Next model into OCI Generative AI, create endpoints for it, and use ...

You can now import the XiaomiMiMo/MiMo-V2.6-Pro-RL model into OCI Generative AI, create endpoints for it, and use ...

You can now import the RedHatAI/NVIDIA-Nemotron-3-Ultra-550B-A55B-FP8-dynamic model into OCI Generative AI, create endpoints for it, and use ...

OCI Generative AI model import now supports more hardware options for the following models.

The Qwen/Qwen3.8-Flash-Next model ...

A new tutorial for OCI Generative AI shows how to route on-demand model requests across OCI regions using ...

Priority processing is now available for the following xAI Grok models in OCI Generative AI:

...

Identity domain browser session authentication, which adds OAuth 2.1 support for hosted deployments, is now generally available in OCI ...

xAI Grok 4.7 is now available on demand in OCI Generative AI. The reasoning model is designed ...

You can now import the zai-org/GLM-5.3-Flash model into OCI Generative AI, create endpoints for it, and use ...

OCI Generative AI now supports model discovery. Use model discovery to find models in a specified region and review ...

OCI Generative AI now supports smart model router for cross-region routing of on-demand inference requests within a user-selected regional ...

You can now import the deepseek-ai/DeepSeek-V4.1-Flash model into OCI Generative AI, create endpoints for it, and use ...

You can now import the google/gemma-4-26B-A4B-it model into OCI Generative AI, create endpoints for it, and use ...

The retirement date for the Cohere Command A Vision (cohere.command-a-vision) and Cohere Embed 4 (cohere.embed-v4.0) ...

You can now import the zai-org/GLM-5.3 model into OCI Generative AI in the UAE Central (Abu Dhabi) region, ...

Cohere Command A Vision (cohere.command-a-vision) and Cohere Embed 4 (cohere.embed-v4.0) are now available in on-demand ...

The on-demand serving mode for Cohere Command A Vision (cohere.command-a-vision) and Cohere Embed 4 (cohere.embed-v4.0) ...

xAI Grok 4.6 is now available on demand in OCI Generative AI. The model is designed for ...

You can now import the Qwen/Qwen3.8-27B and Qwen/Qwen3.5-397B-A17B models into OCI Generative AI, create endpoints for them, ...

You can now import the deepseek-ai/DeepSeek-V4-Pro-0813 model into OCI Generative AI, create endpoints for it, and use ...

You can now import the google/gemma-4-31b-it model into OCI Generative AI, create endpoints for it, and use ...

You can now use the OCI Responses API in OCI Generative AI to call the following imported models: ...

You can now use the target.model.id condition in IAM policies to limit which Generative AI models a group can ...

Enterprise AI NL2SQL now supports model selection, scheduled enrichment, and background processing for SQL-generation requests.

Model selection ...

You can now import the XiaomiMiMo/MiMo-V2.5-Pro model into OCI Generative AI, create endpoints for it, and use ...

You can now import the deepseek-ai/DeepSeek-V4-Flash-0731 model into OCI Generative AI, create endpoints for it, and use ...

You can now import the Qwen/Qwen3.8-2.4T-A95B-FP8 model into OCI Generative AI, create endpoints for it, and use it ...

OCI Generative AI is now available in the US Gov West (Phoenix) region in Oracle US Government Cloud and ...

You can now import the moonshotai/Kimi-K3 model into OCI Generative AI, create endpoints for it, and use it ...

You can now import the following NVIDIA Nemotron 3.5 Lightning models into OCI Generative AI, create endpoints for ...

Google Gemini 2.5 Flash (google.gemini-2.5-flash) is now available in on-demand mode in Brazil East (Sao Paulo) ...

You can now import the following models into OCI Generative AI:

AI Singapore SEA-LION

aisingapore/Qwen-SEA-LION-v4-32B-IT ...

New hardware unit shapes are available for hosting select Cohere and Meta models on dedicated AI clusters in OCI ...

The cohere.command-latest and cohere.command-plus-latest model aliases are no longer available. These aliases resolved to cohere.command-r-08-2024 and cohere.command-r-plus-08-2024, respectively, ...

The OCI Responses API now supports background mode for requests that might take longer than an application wants to ...

On-demand model availability has expanded in OCI Generative AI.

Cohere Embed 4

Cohere ...

H100 multi-node serving is now available for imported models in OCI Generative AI.

You can now deploy deepseek-ai/DeepSeek-V4-Pro ...

OCI Generative AI hosted application endpoints now support OCI IAM authentication.

When you create an application, you ...

New hardware unit shapes are available for hosting select Cohere models on dedicated AI clusters in OCI Generative AI. ...

You can now create private endpoints for imported models in OCI Generative AI. With private endpoints, you ...

The following models are now compatible for import in OCI Generative AI:

MiniMax

MiniMaxAI/MiniMax-M3-MXFP8

The DeepSeek V4 Flash and DeepSeek V4 Pro models are now compatible for import in OCI Generative AI.

The Qwen3 Next 80B A3B Instruct model is now compatible for import in OCI Generative AI.

Qwen3 Next ...

The Google MedGemma 27B Text IT model is now compatible for import in OCI Generative AI.

Google ...

The OpenAI Whisper Large V3 Turbo model is now compatible for import in OCI Generative AI.

The NVIDIA Nemotron 3 Ultra model is now compatible for import in OCI Generative AI, following NVIDIA’s ...

You can now host OpenAI gpt-oss models on dedicated AI clusters with B200 GPU shapes in the UAE Central ...

OCI Generative AI now supports xAI-compatible tools with xAI models through the OCI Responses API. These tools enable ...

You can now use the ApplyGuardrails API in OCI Generative AI to moderate image inputs through the existing content ...

You can now select a specific guardrails version when using the ApplyGuardrails API in OCI Generative AI. Guardrails versions ...

You can now convert text into spoken audio in OCI Generative AI by using Text to Speech with xAI ...

The following models are now compatible for import in OCI Generative AI:

Alibaba Qwen

...

You can now rerank documents with Cohere Rerank 4 in OCI Generative AI. Rerank 4 improves enterprise retrieval workflows ...

OCI Generative AI now supports new features for Cohere Embed 4, including configurable embedding dimensions and mixed ...

OCI Generative AI now is now available in the UAE Central (Abu Dhabi) region. For available models in this ...

Grok 4.3 is the latest reasoning model by xAI, designed for complex, accuracy-critical tasks such as advanced logic, math, ...

OCI Generative AI now offers Enterprise AI agents that speed up the development and deployment of production-grade agentic ...

You can now add Zero Trust Packet Routing (ZPR) security attributes to Generative AI private endpoints. Security attributes are ...

You can now generate and edit images using imported models in OCI Generative AI. For supported models ...

Grok 4.20 (Reasoning and Non-Reasoning) offers industry-leading speed and advanced agentic tool-calling capabilities. It combines a low hallucination rate ...

You can now import the nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16 model into OCI Generative AI, create endpoints for it, and use ...

You can now import the Qwen 3 Embedding models into OCI Generative AI, create endpoints for them, ...

OpenAI gpt-oss models (openai.gpt-oss-20b and openai.gpt-oss-120b) are now available in OCI Generative AI for the dedicated ...

Cohere Embed 4 is now available in OCI Generative AI with on-demand endpoints in the following regions:

You can now create private endpoints for the OCI Generative AI models available in the EU Sovereign Central (Frankfurt) ...

You can now create private endpoints for the OCI Generative AI models available in the UK Gov South (London) ...

You can now enable AI guardrails for content moderation (CM), prompt injection (PI), and personally identifiable information (PII) on ...

You can now enable AI guardrails for content moderation (CM), prompt injection (PI), and personally identifiable information (PII) in ...

You can now access pretrained models through OCI Generative AI private endpoints for the on-demand mode.

An ...

API Keys are now available for OCI Generative AI models.

OCI Generative AI API keys provide ...

Cohere Command A Reasoning and Command A Vision are now available in OCI Generative AI.

Released ...

Summarized thinking traces are now available for Grok Code Fast 1 in the xAI platform for OCI Generative ...

The Google Vertex AI Platform for OCI Generative AI now supports document understanding for the pretrained Gemini 2.5 ...

The Google Vertex AI Platform for OCI Generative AI is now available with on-demand endpoints for the pretrained ...

A new Models by Region page is now available in OCI Generative AI.

This dedicated page ...

OCI Generative AI now supports pretrained OpenAI gpt-oss models designed for reasoning and agentic tasks. Select openai.gpt-oss-120b for ...

You can now import large language models from Hugging Face and OCI Object Storage buckets (in a Hugging ...

We've launched a major update to the Oracle Cloud Console for OCI Generative AI. The new experience ...

The OCI OpenAI package is a Python library that lets you use OpenAI compatible SDKs to call models hosted ...

The Google Vertex AI Platform for OCI Generative AI is now in General Availability (GA) for the pretrained ...

You can now access pretrained models through OCI Generative AI private endpoints.

An OCI Generative AI ...

OCI Generative AI now supports access to the pretrained OpenAI gpt-oss models.

The OpenAI gpt-oss models, designed ...

The Google Vertex AI Platform for OCI Generative AI is now available in Beta. With Beta access, you ...

We're excited to announce the official release of langchain-oci, the dedicated LangChain integration package for Oracle Cloud Infrastructure ...

OCI Generative AI is now available in the EU Sovereign Central (Frankfurt) region through the following API endpoints: ...

OCI Generative AI is now available in the UK Gov South (London) region, through API only.

For API ...

OCI Generative AI is now available in the Saudi Arabia Central (Riyadh) region. For more information see Regions ...

OCI Generative AI is now available in the US East (Ashburn) region. For more information see Regions with ...

OCI Generative AI is now available in the US West (Phoenix) region. For more information see Regions with ...

The xAI platform for OCI Generative AI now supports access to the pretrained Grok 4 model in Generative ...

OCI Generative AI hosts the following models in the India South (Hyderabad) region.

Chat Models

...

OCI Generative AI now supports model aliases for applications and client code that use models in the on-demand ...

OCI Generative AI has added Console messages, API date parameters, and OCI notification announcements for model deprecation and ...

OCI Generative AI is now available in the India South (Hyderabad) region. For more information see Regions with ...

OCI Generative AI has added support for Cohere Embed 4. This model can process text or images ...

OCI Generative AI models are now supported in LangChain4j, the Java version of LangChain. This integration expands ...

The GenerateText and SummarizeText APIs are now deprecated. From 26th June 2026 onwards the deprecated APIs will be unavailable. ...

The xAI platform for OCI Generative AI is in General Availability and you can now access the pretrained ...

The xAI platform for OCI Generative AI is now available in Beta. With Beta access, you can now ...

OCI Generative AI now supports Meta Llama 4 models, Scout and Maverick, on Oracle Cloud Infrastructure (OCI) Generative ...

OCI Generative AI now supports the following Cohere models on Oracle Cloud Infrastructure (OCI) Generative AI service. ...

OCI Generative AI is now available in the UAE East (Dubai) region. For more information see Regions with ...

You can now add content moderation to OCI Generative AI model endpoints.

Content moderation feature aims to ...

You can now add an optional safety mode to OCI Generative AI Cohere models during a chat session. ...

OCI Generative AI now supports fine-tuning Meta's pretrained 70 billion-parameter Llama 3.3 instruct model. This text-only model delivers ...

OCI Generative AI now supports TypeScript code samples in the OCI Generative AI playground experience.

OCI ...

OCI Generative AI now supports Meta's 70 billion-parameter Llama 3.3 instruct model. This text-only model delivers better performance ...

OCI Generative AI has added tool support to Llama models, enabling them to integrate with external tools and ...

OCI Generative AI has released a new FP8 quantized version of the Meta Llama 3.1 405B model, with ...

OCI Generative AI now supports a seed parameter in the OCI Generative AI playground, CLI, and API. The ...

OCI Generative AI now supports LangChain and LlamaIndex code samples in the OCI Generative AI playground experience. ...

OCI Generative AI now supports Meta's 90 billion-parameter and 11 billion-parameter Llama 3.2 vision models in all its ...

OCI Generative AI is now available in the Japan Central (Osaka) region. For more information see Regions with ...

OCI Generative AI now supports Meta's 90 billion-parameter and 11 billion-parameter Llama 3.2 vision models. These models offer ...

OCI Generative AI now supports the latest updates to Cohere's Command R and Command R+ models (08-2024). These ...

OCI Generative AI now supports the pretrained Meta Llama 3.1 70 billion-parameter and 405 billion-parameter large language models. ...

OCI Generative AI is now available in the Brazil East (Sao Paulo) region. For more information see Regions ...

OCI Generative AI now supports fine-tuning the pretrained Cohere Command R model. Command R is an instruction-following conversational model used ...

OCI Generative AI now is now available in the UK South (London) region. For more information see Regions with Generative ...

OCI Generative AI now supports the pretrained Cohere Command R + model. Command R+ is an enhanced version of the ...

OCI Generative AI now supports a chat API to provide an AI chatbot experience for supported Generative AI models. ...

We are excited to announce that OCI Generative AI now supports the Meta Llama 3 70B model, a cutting-edge large ...

OCI Generative AI now supports the pretrained Cohere Command R model. Command R is an instruction-following conversational model that performs ...

OCI Generative AI now is now available in the Germany Central (Frankfurt) region. For more information see Regions with Generative ...

We are thrilled to announce the General Availability (GA) of OCI Generative AI service in the ORD (us-chicago-1) region.

Generative AI is now available in Beta.

OCI Generative AI (Beta) is a fully managed service that provides a ...