Ask HN: How do GitHub Copilot and Anthropic Claude compare in cost?

At the company I work at the developers all have a GitHub Copilot subscription. Copilot allows them to use several models, including Claude. From the usage I can actually see that Claude Sonnet and Opus are the most commonly used models. This got me thinking: would it be more efficient to switch to Claude completely? It is difficult to compare the GitHub credits with the Claude tokens, but does someone have experience using both? Is the difference significant?