FAQ - Qodo Merge - Can Qodo Merge handle large PRs?

Generally speaking, there are two main solutions to handle code suggestions for large* PRs in Qodo Merge:

1. Enable the 'extended' mode for the improve tool by using /improve --extended. This mode divides the PR into chunks and processes each chunk separately, allowing it to handle larger PRs without compression. Note that for large PRs, this may result in multiple model calls.
2. Use a model with a larger context, like GPT-32K or Claude-100K. This can be configured in your configuration file and will apply to all tools.

You can control the LLM model Qodo Merge is using through your configuration file (please see full details here on the configuration file) by setting (setting the model to gpt-4o as example):

[config]
model="gpt-4o"
model_turbo = "gpt-4o"
fallback_models=["gpt-4o"]

* We define “large PR” as anything greater than 3000 characters