What's New on DigitalOcean Gradient™ AI Platform

Updated: February 25, 2026
17 min read

Welcome to What’s New on DigitalOcean Gradient™ AI Platform—your weekly roundup of the latest updates for the Gradient AI Platform.

Week of February 23rd

Product Update Now Available: OpenAI GPT-5.3-Codex for Frontier Coding and Agentic Execution

OpenAI GPT-5.3-Codex is now live on DigitalOcean’s Agentic Inference Cloud. Build advanced coding agents, operate across complex codebases, and run multi-step workflows via Gradient™ AI Serverless Inference, with security-hardened defaults, usage-based pricing, and fully managed infrastructure. GPT-5.3-Codex combines highly advanced coding performance with stronger terminal and computer-use capabilities, enabling long-running, tool-driven workflows and real-world professional tasks. The result: fewer iterations, faster execution, and more production-ready outputs.

What’s new from OpenAI:

Stronger multi-step reasoning for long-running agent workflows
Improved computer use with advanced terminal fluency and desktop task execution
Expanded support for knowledge work across presentations, spreadsheets, research, and software lifecycle tasks
State-of-the-art results on SWE-Bench Pro
77.3% accuracy on Terminal-Bench 2.0

Start using GPT-5.3-Codex today through the DigitalOcean API or directly in the Cloud Console. To learn more about the model and how to access it via DigitalOcean, check out our product documentation.

Week of February 17th

Product Update Now Available: Claude Sonnet 4.6, Anthropic’s most capable Sonnet-class model yet

Claude Sonnet 4.6 from Anthropic is now live on DigitalOcean’s Agentic Inference Cloud. Build advanced agents and tackle complex codebases via Gradient™ AI Serverless Inference with security-hardened defaults, usage-based pricing, and no infrastructure to manage. Sonnet 4.6 delivers major upgrades across coding, computer use, long-context reasoning, agent planning, and knowledge work to help teams ship faster with fewer iterations.

What’s new from Anthropic:

Stronger coding performance with better instruction following and fewer hallucinations
Improved computer use, with early users reporting near human-level capability on tasks like navigating complex spreadsheets and completing multi-step web workflows
Agentic financial and office task improvements, bringing Opus-class performance to more practical workloads
1M token context window (beta) for reasoning across entire codebases or large documents

You can start using Sonnet 4.6 today through the DigitalOcean API or directly in the Cloud Console. To learn more about the model and how to access it via DigitalOcean, check out our product documentation.

Product Update Introducing DigitalOcean Managed Keys for Third-Party Models

DigitalOcean Managed Keys let you build and deploy AI agents without BYOK or external setup, with all usage billed directly to your DigitalOcean account. Frictionless access, unified billing, and native key management simplify operations so you can move from development to production faster. Learn more about our available models or create a model access key now ->

Week of February 9th

Product Update New Models from OpenAI, Anthropic, and fal

Gradient AI Platform now supports top models across text, code, image, and audio generation, including OpenAI, Anthropic, and fal. All models run inside your existing DigitalOcean environment, with one control plane for applications, data, networking, and inference.

No separate vendor accounts or model contracts. Usage is billed predictably with your existing DigitalOcean services, with managed inference and safe defaults so you can start building immediately.

Access models via API or the DigitalOcean Cloud Console (media generation models via API only).

Available models include:

Text & Code: OpenAI GPT-5.2, GPT-5.2 Pro, GPT-5.1-Codex-Max; Anthropic Haiku 4.5, and Claude Opus 4.6
Image: fal Stable Diffusion XL Fast; fal FLUX.1 (schnell)
Audio: fal Stable Audio; ElevenLabs TTS Multilingual v2

Explore available models →

For more directions on how to use the image and audio models, check out our product documentation ->

Week of February 5th

Product Update Now Available: Anthropic Claude Opus 4.6

Claude Opus 4.6 is now available on the DigitalOcean Gradient™ AI Platform via Serverless Inference, giving teams access to Anthropic’s most capable model on a platform built to run inference reliably at scale. With massive 1M-token context, adaptive reasoning, and advanced agentic coding, Claude Opus 4.6 enables teams to analyze huge datasets, refactor entire codebases, and generate high-quality outputs in a single pass. It’s also optimized for everyday knowledge work, including reports, spreadsheets, and presentations.

Use Opus 4.6 today via the DigitalOcean API or in the Cloud Console. Want to learn more about Opus 4.6 and accessing it via DigitalOcean? Check out our article here.

Week of January 26

Product Update Now Available: Prompt caching for these Anthropic models

Use prompt caching with serverless inference chat completion to cache context and use it in future requests. When you send a request, if part of the input is already cached, you are charged a lower price for those cached tokens, and at the standard price for the remaining input tokens. This significantly reduces the cost for inference.

We support prompt caching for the following Anthropic models:

Claude Sonnet 4
Claude Opus 4
Claude Opus 4.1
Claude Sonnet 4.5
Claude Opus 4.5

Week of December 15

Product UpdateIntroducing the Gradient™ AI Agent Development Kit

The ADK, now in public preview, is a code-first SDK for building and operating production-ready AI agents across popular Python frameworks, including LangGraph, LangChain, CrewAI, and AutoGen. Built-in tracing, evaluations, and Knowledge Base support make it easier to move from local development to production. Enable the ADK public preview now ->

Product Update New Anthropic models now available on the Gradient™ AI Platform

Claude Opus 4.1, Opus 4.5, and Sonnet 4.5 are now available on the DigitalOcean Gradient™ AI Platform for Serverless Inference and the Agent Development Kit, giving developers more powerful options for building and scaling AI applications. For details, see the Available Models page ->

Product Update DigitalOcean Gradient™ AI Platform Knowledge Base Enhancements

The enhanced DigitalOcean Gradient™ AI Knowledge Base, now in public preview, lets developers create, manage, and query knowledge bases entirely in code, giving full control over ingestion, chunking, embedding, and retrieval. With direct API access, flexible ingestion, and citation-backed search, the enhancements makes it easier to power accurate, production-ready RAG workflows. Enable the KB enhancements public preview now ->

Product Update Upgraded Traces & Log Stream Insights Now Live

We’ve enhanced Log Stream Insights with a more prominent Insights button and the ability to navigate directly to affected spans, making it easier to investigate issues. Plus, our improved Insights model delivers richer, more accurate trace analysis at a fraction of the cost—reducing output token pricing from $20 to just $4.40.

image alt text

Week of December 1

Product Update Evaluate your AI agents more effectively and faster We’ve redesigned agent evaluations on the DigitalOcean Gradient™ AI Platform to make testing AI agents faster, clearer, and easier to adopt. The new experience introduces goal-oriented metric groups, example datasets, clearer error messages, and deep trace integration to help developers diagnose and improve performance quickly. Start evaluating your agents today to identify issues, optimize behavior, and build safer, more reliable, production-ready systems. Learn more in the announcement blog ->

Week of October 27

Product Update Build smarter knowledge bases with Sitemap support

You can now add a sitemap URL (in .xml format) to crawl a specific list of pages all at once, rather than adding URLs individually. Simply enter your sitemap (e.g., [docs.digitalocean.com/sitemap.xml](http://docs.digitalocean.com/sitemap.xml)), and optionally include header and footer links or embedded media for more complete indexing. Learn more in our product documentation →

image alt text

Week of October 20

Product Update New: Image and audio generation models from fal now in public preview on DigitalOcean Gradient™ AI Platform

You can now generate high-quality images and audio on the Gradient AI Platform using fal’s cutting-edge models. fal specializes in delivering fast, production-ready generative models that make multimodal creation simple and scalable. These new models are available in public preview through Serverless Inference—no infrastructure setup needed. Check out the announcement blog for instructions on how to get started.

Week of September 29

Product Update Image Model Support

Generate images from text prompts using the Serverless Inference API powered by OpenAI’s gpt-image-1 model. Billing, API access, and authentication are fully integrated for developers and enterprises.

Product Update Knowledge Base Auto-Indexing

Keep agents’ knowledge current automatically by fetching and re-indexing new or updated documents from sources like Google Drive, S3, or Dropbox.

Product Update VPC Integration

Securely run agents and indexing jobs with private network access, cross-VPC connectivity, and visibility into outbound IPs for enterprise deployments.

Product Update Gradient AI Agent Development Kit (ADK) – Private Preview

Build, test, and deploy AI agent workflows entirely in code, with model integration, custom connectors, and workflow tracing directly in your workspace.

Product Update Gradient AI Genie – Private Preview

Interact with and orchestrate multi-agent systems from inside VS Code, get workflow guidance, and automate agent tasks without leaving your IDE.

Week of September 22

Product Update Knowledge Base Activity Logs

See exactly what happens when you add data to a knowledge base. Logs show real-time job status, history, and per-source errors so you can debug quickly.

Product Update Agent Templates

Start faster by cloning and customizing code-first agent templates for tasks like SQL queries, documentation chatbots, fact-checking, and Twilio messaging.

Tutorial Simplified Testing with Gradient™ AI SQL Agent Template

This tutorial demonstrates how to extend DigitalOcean’s Gradient™ AI SQL Agent Template with synthetic data generation, enabling safe, realistic mock datasets, multi-format outputs, and natural language–driven SQL queries for testing and development.

Product Update n8n Integration

Add Serverless Inference to n8n workflows without coding, with multi-role chat, multiple model access, and customizable parameters.

Product Update Enhanced Guardrails

Protect agents from malicious inputs with improved jailbreak detection, broader coverage, and real-time input blocking.

Week of September 8

Product Update New GPT-5 models

GPT-5 mini model and GPT-5 nano models are now available on DigtialOcean Gradient AI Platform. For more information, see the Available Models page.**

Week of August 25th

Product Update Developer Deep Dive: Code-First Gradient AI Platform Workflows

On August 27, we hosted a Developer Deep Dive session showcasing code-first workflows on the Gradient AI Platform. The live walkthrough covered deploying models and agents with the new SDK, integrating with LangChain for orchestration, using the DigitalOcean MCP Server for natural language → API automation, and a behind-the-scenes look at how we built and shipped a SaaS app on Gradient. Watch the recording:

Tutorial Build an AI Customer Support Agent with GPT-OSS

Learn how to build an AI-powered customer support agent using the gpt-oss-120B model on the Gradient AI Platform, with steps for creating knowledge bases, grounding responses, and deploying your agent.

Tutorial Automated Invoicing with AI: Stripe + Resend + DigitalOcean

Learn how to use serverless inferencing on Gradient AI Platform to fetch data from Stripe and automatically generate invoices, showcasing a seamless AI-powered billing workflow.

Week of August 18th

Product Update Introducing langchain-gradient: Seamless LangChain Integration with Gradient AI Platform

Introducing langchain-gradient, our new open-source integration that lets LangChain pipelines plug directly into Gradient AI Platform—so you can orchestrate LLM-powered workflows, access multiple models via serverless inference, and scale AI apps with just a few lines of code. Check out the integration now ->

Product Update New OpenAI OSS models available: gpt-oss-120b and gpt-oss-20b

On the Gradient AI Platform, you can now access the gpt-oss-120b and gpt-oss-20b open-source models for inference. See all available models →

Tutorial SeaNotes — One-Click Deploy SaaS Starter Kit

Learn how to use SeaNotes, a production-ready, open-source SaaS starter kit built with Next.js and DigitalOcean services (Spaces, Gradient, PostgreSQL, App Platform), plus Stripe and Resend. This video walks through running it locally, deploying to App Platform, and leveraging built-in features like auth, billing, email, file uploads, and AI with Gradient.

Week of August 11th

Product Update New OpenAI models available: GPT-5 and o1

On the Gradient AI Platform, you can now access the latest OpenAI models, including GPT-5 and o1, for inference. See all available models →

Week of August 4th

Webinar Recording Webinar: Building AI Agents on Gradient AI Platform

On August 6, we hosted a live webinar on how to build and deploy AI agents using the Gradient™ AI Platform. Nearly 600 attendees joined to see demos, best practices, and get hands-on guidance. Watch the recording:

Week of July 28th

Product Update Add Dropbox folders as a data source

On Gradient AI Platform, you can now add a Dropbox folder as a data source to your knowledge bases. This allows you to index and use files stored in your Dropbox account within your knowledge base.

Week of July 21st

Product Update Log Stream Insights on Gradient AI Platform

DigitalOcean GradientAI is a unified AI cloud platform that combines GPU infrastructure, intelligent agent development, and prebuilt applications to help developers seamlessly build, deploy, and scale AI solutions from prototype to production.

Product Update Gradient AI Platform SDK Now Available in Public Preview

The official DigitalOcean Platform SDK is now in Public Preview. You can use the SDK to manage Gradient AI Platform resources, including knowledge bases and generative AI agents, from Python applications.

Week of July 7th

Product Update Introducing DigitalOcean Gradient™ AI

Product Update Gradient AI Platform is now GA

Gradient AI Platform is now generally available, offering developers a fully managed environment to build, scale, and deploy AI-powered applications and agents—complete with integrated tools for data integration, evaluation, observability, and serverless access to top LLMs—all without managing infrastructure.

Tutorial How I built a smart travel app using the Gradient AI Platform

This tutorial shows how to build an AI-powered travel app called Nomado using DigitalOcean’s Gradient AI Platform. It walks through using serverless inference to generate personalized itineraries, function routing for real-time flight info, and agent routing to power a chatbot for accommodation and cultural tips—all without managing infrastructure.

Tutorial Using DigitalOcean’s Serverless Inference (and Agents) with the OpenAI SDK

This tutorial shows how to use the OpenAI SDK with DigitalOcean’s Serverless Inference and Agents by simply changing the base URL—letting you run powerful LLMs or connect them to your own data without rewriting code, while supporting streaming, retrieval metadata, and seamless backend swapping.

Week of June 30th

Product Update Agent Traceability and Conversation Logs are now available

Agent traceability: New debugging tools show you how each request flows through your agent, including routing decisions, token counts, and system messages. This helps you pinpoint issues, improve accuracy, and optimize performance in production.

Customer conversation logs: Capture and store agent interactions across your app, playground, or widget. With role-based access controls (RBAC) and data retention controls, you can export and analyze logs for analyzing user behavior, training better agents, or supporting audit and compliance needs.

Week of June 17th

Webinar Recording Integrate AI with Serverless Inference on DigitalOcean

Watch this recorded session to learn how DigitalOcean’s serverless inference makes it easy to integrate AI into your app—no setup, no idle costs, and auto-scaling included. You’ll see real-world demos using models from OpenAI, Anthropic, Meta, and more, all accessible through a single API.

Week of June 9th

Product Update Introducing Serverless Inference on the GenAI Platform

Serverless inference is now available on the GenAI Platform, allowing you to get direct responses from foundation models without creating an agent. A single API endpoint provides access to hosted models from providers like OpenAI, Anthropic, Meta, and others. You can authenticate requests with model access keys, which provide secure access to all available models.

Tutorial Serverless Inference with the DigitalOcean Gradient AI Platform

This tutorial walks you through how to set up and run serverless inference on the DigitalOcean Gradient AI Platform using the API or Cloud Console, showing how to generate text with a pre-trained model in Python and highlighting real-world use cases like event-driven apps, scalable backends, and data processing.

Week of May 19th

Tutorial Effective Strategies for Sending Data to AI Agents

This tutorial provides a comprehensive guide to preparing, formatting, and transmitting data to GenAI agents, covering best practices for handling structured and unstructured inputs, managing large datasets, enabling real-time streaming, and optimizing performance to ensure efficient, scalable AI integration with platforms like Gradient AI Platform.

Week of May 5th

Tutorial Knowledge Base Citations, Agent Versioning, and Insights on the Gradient AI Platform

This tutorial shows you how to use new Gradient AI Platform features—Knowledge Base Citations, Agent Versioning, and Agent Insights—to build more transparent, trackable, and cost-aware AI agents, with step-by-step guidance on setting up a knowledge base, tracing model outputs, managing agent iterations, and monitoring usage at scale.

Week of April 28th

Product Update New: Agent Versioning, Agent Insights, and Knowledge Base Citations

DigitalOcean’s GenAI Platform now supports enhanced agent development with agent insights, agent versioning, and knowledge base citations. These updates make it easier than ever to build transparent, auditable, and high-performance AI agents for real-world applications.

Week of April 21st

Product Update Improvements to our Agent Studio

We’ve improved the Agent Playground and agent API experience. You can now see the knowledge bases, functions, and guardrails that is used by your GenAI Platform agent to generate a response.

Product Update Runtime Logs

You can now view the runtime logs for your agents on GenAI Platform. The logs display the events that occur during an agent’s execution, such as the knowledge bases and functions accessed to generate a response.

Product Update Gradient AI Platform Agent Templates

You can now create Gradient AI Platform agents from templates. Templates have predefined agent instructions and foundation models that allow you to quickly create an agent for certain use cases.

Tutorial Optimizing OpenAI’s GPT with Gradient AI Platform

This tutorial walks you through how to use Gradient AI Platform to build an AI agent powered by OpenAI’s GPT-4o or GPT-4o mini, showing step-by-step how to set up a knowledge base, integrate your data with RAG, and deploy a reasoning-capable agent using your own OpenAI API key.

Week of April 14th

Product Update Smarter Knowledge Bases for Smarter AI Agents

Now available on DigitalOcean’s GenAI Platform: enhanced knowledge base features including web crawling, custom crawl rules, and one-click reindexing, making it easier to keep your AI agents up to date with real-time, context-rich information.

Product Update Claude Sonnet 3.7 Model now available on the GenAI Platform

Claude Sonnet 3.7 is now live on the GenAI Platform. Check out the available models page for details and usage info.

Webinar Recording Sail to Success: GenAI Platform Knowledge Base Best Practices

Hear from experts at DigitalOcean and Vijil about building a powerful Knowledge Base for AI agents. Learn to integrate documentation, FAQs, and custom data, tackle common challenges, and apply best practices. Discover how new features like web crawling and faster indexing can enhance performance. Whether you’re starting or optimizing your AI workflow, this session will equip you with the insights and tools to maximize GenAI’s potential.

Week of April 7th

Product Update Expanding the GenAI Platform: Now supporting OpenAI models

DigitalOcean’s GenAI Platform now supports OpenAI models, allowing users to integrate GPT-4o, o1, and o3-mini into their AI agents using their own OpenAI API keys. This update provides greater flexibility for building AI applications with OpenAI’s advanced models, offering users the ability to test and deploy models in production seamlessly.

Tutorial Building a Slack AI Chatbot with DigitalOcean GenAI Platform

In this tutorial, learn how to build a Slack AI chatbot using DigitalOcean’s GenAI Platform to automate team interactions, such as summarizing conversations, answering questions, and responding to mentions or direct messages, by integrating large language models directly into your Slack workspace.

Tutorial Build an AI Agent to Automate Document Analysis with GenAI

This tutorial demonstrates how to create an AI agent using the GenAI Platform to automate the analysis of security questionnaires, including setting up a private API endpoint, developing a Streamlit application to process Excel files, and deploying the solution on DigitalOcean’s App Platform.

Week of March 31st

Tutorial Build Real-Time AI Agents with GenAI and Serverless Functions

This tutorial demonstrates how to build real-time AI agents on DigitalOcean’s GenAI Platform by integrating serverless functions, enabling dynamic data retrieval and processing.

Video Build an AI Agent to Automate Security Questionnaires

Sick of answering the same security and compliance questions over and over? In this video, learn how you can build a practical AI-powered app that analyzes your company’s compliance documents and responds to vendor questionnaires using Generative AI and RAG.

Week of March 24th

Product Update Improved Web Crawling for Knowledge Bases

The limit of URLs that can be crawled by the GenAI web crawler for data sources has no been increased from 1,000 to 5,500.

Tutorial Beyond Vectors - Knowledge Graphs & RAG Using GenAI

This tutorial demonstrates how to enhance Retrieval-Augmented Generation (RAG) systems by integrating knowledge graphs, enabling AI models to understand complex relationships within data for more accurate responses.

Week of March 17th

Tutorial Building an AI Article Reviewer using GenAI Platform and GitHub

This tutorial demonstrates how to create an AI article reviewer by integrating DigitalOcean’s GenAI Platform with GitHub Actions, enabling automated, style-consistent code reviews.

Tutorial Building Real-Time Kubernetes AI Agent with GenAI Platform

In this article, explore how to build a real-time AI agent for Kubernetes using DigitalOcean’s GenAI Platform, enabling automated generation, validation, and troubleshooting of Kubernetes manifests.

Tutorial DeepSeek-R1 vs. Llama 3.3 (70B): AI Chatbot on GenAI

This tutorial compares the DeepSeek-R1 and Llama 3.3 (70B) AI chatbots deployed on DigitalOcean’s GenAI Platform, evaluating their performance on startup-related queries to help you choose the best model for your needs.

Video GenAI Platform: Using Function Calling to retrieve real-time data

One of the biggest challenges with AI models is their lack of access to real-time data from external sources. But what if you could create an AI-powered agent that connects to APIs for up-to-date answers? Watch the demo to see how simple it is to integrate external data into your AI applications.

Week of March 10th

Tutorial Build an AI Agent or Chatbot with GenAI Platform Integration

The guide provides you with step-by-step instructions on creating a customizable AI agent or chatbot using DigitalOcean’s GenAI Platform, covering model selection, knowledge base integration, and deployment.

Week of February 24th

Product Update Web Crawling for Knowledge Bases

You can now add a website as a data source for your GenAI knowledge bases. Providing a seed URL and a scope allows GenAI to crawl the website for information, index it, and use it to train your agent models.

Product Update Sensitive Data Detection for Guardrails

With sensitive data detection guardrails, the agent automatically identifies and anonymizes sensitive data—such as credit card numbers, personal identifiable information, and location data—with fully customizable category settings.

Week of February 17th

Tutorial How to Deploy DeepSeek R1 on DigitalOcean

This article outlines three ways you can deploy the DeepSeek R1 LLM on DigitalOcean—via GenAI Platform, Hugging Face integration, or GPU Droplets with Ollama.

Tutorial Getting Started with the GenAI Platform

Want to get started with AI on DigitalOcean? This guide walks you through using the GenAI Platform to create a knowledge base, add guardrails, and deploy your own AI agents.

Week of February 10th

Product Update Anthropic and DeepSeek Join the GenAI Platform Model Library

DigitalOcean’s GenAI Platform now supports Anthropic’s Claude 3.5 and DeepSeek R1 models, giving you more powerful, flexible options for building AI agents—ideal for everything from real-time chat to complex problem-solving. With serverless model hosting, easy integration, and pay-as-you-go pricing, you can deploy advanced AI applications without worrying about infrastructure. Try the new models ->

Video Building a Market Research Assistant with GenAI Platform, DeepSeek R1, and CrewAI

In this video, learn how you can build a Market Research Assistant using powerful tools like DigitalOcean’s GenAI Platform, DeepSeek R1, and CrewAI.

Week of February 3rd

Video Building a Multi-Agent Chatbot with Function Routing on Gradient AI Platform

In this demo, learn how to build a multi-agent chatbot solution on the Gradient AI Platform using function routing.

Week of January 27th

Product Update Introducing the DigitalOcean Gradient™ AI Platform: Simplifying AI Development for All

DigitalOcean Gradient™ AI Platform is an all-in-one solution that enables you to quickly build and scale AI agents, leveraging advanced Generative AI models from providers like Meta and Mistral AI. Designed for both seasoned developers and newcomers, the new platform offers features such as retrieval-augmented generation (RAG), function routing, agent customization, and guardrails to facilitate the creation of secure, context-aware applications. Build an agent ->

Product Update Knowledge Base Improvements

You can now add files from your local storage as a data source to your knowledge base.

Tutorial The Secret Sauce to a Winning Dataset for Gradient AI - Quality Over Quantity

This article highlights the importance of prioritizing data quality over quantity when building datasets for Retrieval-Augmented Generation (RAG) applications, offering insights into best practices for gathering and evaluating data sources.

Tutorial From Knowledge Gaps to Expert Agents - Unlocking the Advanced Use Cases of RAG with the DigitalOcean GenAI Platform

This article dives into the advanced features of DigitalOcean’s Gradient AI Platform, specifically Function Calling and Routing Multiple Agents, demonstrating how these capabilities enable developers to create customized, scalable AI systems by integrating external functions and managing specialized agents for complex tasks.