Skip to main content
OpenHands is an open-source, model-agnostic AI agent platform for software development that executes real engineering work end-to-end across your codebase. Follow the steps below to connect OpenHands with a model hosted on SambaCloud.

​
Prerequisites

Before starting, ensure you have:

​
Setup

1

Open LLM settings in OpenHands

Log in to OpenHands and navigate to Settings from the left-hand panel.Open the Language Model (LLM) section and click Advanced to reveal the custom model configuration fields.
2

Add SambaNova as a custom model

Enter your SambaNova model using the sambanova/ prefix.Example:
sambanova/MiniMax-M2.7
Verify the following settings:
  • Custom Model: sambanova/<model-name>
  • Base URL: https://api.sambanova.ai/v1
  • API Key: Your SambaNova API key
Click Save Changes after completing the configuration.
3

(Optional) Configure an organization-wide default

If you are an OpenHands Cloud organization administrator, you can configure SambaNova as the default model for your team.Navigate to Settings → LLM Profiles and create a new profile using the same:
  • Custom Model
  • Base URL
  • API Key
Set the profile as the organization default. Team members can still override the setting with personal configurations if needed.
4

Start using SambaNova models

Create a new conversation in OpenHands and assign a software engineering task, such as:
  • Find and fix the bug in my Python script
  • Write unit tests for this function
  • Refactor this code to use async/await
OpenHands will use your selected SambaNova model to plan, write code, execute commands, and iterate on solutions.

​
Self-hosted and Docker configuration

If you’re running OpenHands with Docker, configure SambaNova using environment variables: 
docker run \
  -e LLM_MODEL="sambanova/MiniMax-M2.7" \
  -e LLM_BASE_URL="https://api.sambanova.ai/v1" \
  -e LLM_API_KEY="<your-sambanova-api-key>" \
  docker.all-hands.dev/all-hands-ai/openhands:latest

​
Additional resources