Announcing OpenAI gpt-oss Models on the DigitalOcean Gradient™ AI Platform

Announcing OpenAI gpt-oss Models on the DigitalOcean Gradient™ AI Platform | DigitalOcean

© 2026 DigitalOcean, LLC. Sitemap .

Dark mode is coming soon. Product updates Announcing OpenAI gpt-oss Models on the DigitalOcean Gradient™ AI Platform

By Grace Morgan

Updated: August 20, 2025 2 min read

<- Back to blog home

OpenAI’s first open-source GPT models (20b and 120b) are now available on the Gradient AI Platform . This launch brings even more flexibility and choice to developers building AI-powered applications, whether you’re starting with a quick prototype or scaling a production agent.

What’s new

Open-source GPT models: Access gpt-oss 20b and 120b directly on the Gradient AI Platform.

Code + UI support: Call the models through our Serverless Inference API or select them in the Gradient dashboard when creating agents, or try them out in the model playground.

Integrated experience: Unified billing, observability, and traceability built into the platform—no need to stitch together multiple vendors, billing, or monitoring tools.

How to get started

With code: Call the models directly through our Serverless Inference API.

python

import os import sys from gradient import Gradient

model_access_key = os.environ.get( "GRADIENT_MODEL_ACCESS_KEY" )

if not model_access_key: sys.stderr.write( "Error: GRADIENT_MODEL_ACCESS_KEY environment variable is not set. \n " ) sys.exit(1)

inference_client = Gradient(model_access_key=model_access_key)

inference_response = inference_client.chat.completions.create( messages=[ { "role": "user", "content": "Write a product description for an eco-friendly water bottle.", } ], model="openai-gpt-oss-20b", )

print(inference_response.choices[0].message.content)

In the UI: Head to Agent Creation and select gpt-oss 20b or 120b from the model dropdown.

What you can build

Customer support agents trained with your knowledge base

Content generation tools with cost-optimized model choices

AI-powered apps that balance performance with affordability

This launch marks another step toward making Gradient AI Platform the simplest, most flexible way to build real, production-ready AI applications.

Deploy now

For code-first developers:

👉 Deploy gpt-oss via API: Spin up the 20b or 120b models directly through the Gradient Serverless Inference API and start building in just a few lines of code.

For UI users:

👉 Build with gpt-oss in console: Create an agent in the Gradient AI Platform, select gpt-oss from the model dropdown, and deploy instantly—no code required.

About the author

Grace Morgan Author

See author profile

See author profile

Share

Product Updates

Start building today From GPU-powered inference and Kubernetes to managed databases and storage, get everything you need to build, scale, and deploy intelligent applications. Sign up

Related Articles

Product updates Model Evaluations: Prove Your Routing Policy Actually Works

Sathish Jothikumar

June 4, 2026 7 min read

Read more

Product updates Powering the Inference Era: Inside the DigitalOcean Data & Learning Layer

Zach Peirce

June 3, 2026 5 min read

Read more

Product updates OpenCode Now Supports DigitalOcean Inference Router for Intelligent Model Routing

Musa Malik May 28, 2026 3 min read

Read more