Code of Conduct
What article on docs.github.com is affected?
https://docs.github.com/en/copilot/reference/copilot-billing/models-and-pricing
What part(s) of the article would you like to see updated?
OpenAI models (specifically GPT-5.6) on https://docs.github.com/en/copilot/reference/copilot-billing/models-and-pricing do not have information about cache costs.
But when tested, they do cost for cache writes, please update the documentation. I assume the same is true for older models also.
Additional information
Each request to GPT-5.6 Sol, Terra and Luna costs more than one can see from the cost table - the missing piece is cache write. With that included (with pricing for e.g. Sol as $6.25 (just like in OpenAI API pricing)) costs are aligned.
Code of Conduct
What article on docs.github.com is affected?
https://docs.github.com/en/copilot/reference/copilot-billing/models-and-pricing
What part(s) of the article would you like to see updated?
OpenAI models (specifically GPT-5.6) on https://docs.github.com/en/copilot/reference/copilot-billing/models-and-pricing do not have information about cache costs.
But when tested, they do cost for cache writes, please update the documentation. I assume the same is true for older models also.
Additional information
Each request to GPT-5.6 Sol, Terra and Luna costs more than one can see from the cost table - the missing piece is cache write. With that included (with pricing for e.g. Sol as $6.25 (just like in OpenAI API pricing)) costs are aligned.