This lesson's interactive features are locked, please to keep using them

OpenCode Console

OpenCode Console is a suite of hosted services offered by the same team that develops the OpenCode agent. The console gives you access to a large catalog of AI coding models, some of which are free, which means you can complete this course without paying anything.

You don't have to authenticate to Console to use its free models – but we recommend it, as there are certain features that do require auth.

Of course, the OpenCode agent can also connect directly to other providers. In this course, if you don't already have a preferred provider, we recommend using OpenCode Console with a free model. If you already have a ChatGPT, Google Gemini, or other provider account, we recommend just using that.

The Connect Command

Using the /connect command inside an OpenCode session connects to a provider like OpenCode Console or ChatGPT:

/connect

The Models Command

The /models command opens the model selector, so you can search for a model and switch to it.

DeepSeek V4.1 Flash

If you don't already have a preferred provider – for example, a Claude account with an API key, or a ChatGPT account with a paid subscription – we recommend DeepSeek V4.1 Flash Free, provided through OpenCode Console.

The DeepSeek V4.1 Flash Free endpoint is available for a limited time. OpenCode states that the provider retains no request data and doesn't use it for model training. Still, don't send private information or credentials through this.

Assignment

Select a model. Using OpenCode Console is optional, but you need to use something.

Run /models in an OpenCode session in the project repo.
Search for and select DeepSeek V4.1 Flash Free and choose the high thinking variant. (Or select your model of choice.)
Tell OpenCode to run opencode models and confirm that you see a list of <provider>/<model> entries.

Have OpenCode submit the CLI tests from the project root. Again, you can copy-paste the prompt from the test panel.