🏷️ Do you know which model your “gpt-5.6” calls are actually billed as?
There are two lines on OpenAI’s GPT-5.6 model page that are easy to miss if you call the model by its alias.
The first: the plain gpt-5.6 alias routes to GPT-5.6 Sol. If your code asks for gpt-5.6 because it looked like the sensible default, you’re paying Sol rates: $4 per million input tokens, $20 per million output.
The second: Sol’s pricing is labelled promotional, “available at least through November 21, 2026”. The page doesn’t say what it becomes after that.
Take an ordinary feature: 100,000 requests a month at 1,200 tokens in and 300 out. That’s $480 of input plus $600 of output, so $1,080 a month, on a price that’s only promised until the 21st.
Is your model name pinned somewhere you’d notice if the price under it moved, or would the invoice be the first place you see it?
Replies
Not from the invoice. I watch API usage all the time, and when it crosses a threshold I get a notification on my phone. So if the price under an alias moved, I'd hear about it from my phone before the bill. Good catch on the end date, though.
Pinned, and in config rather than code. My backend names each music model in an environment value, one for the standard tier and one for the fallback, and the price of each version sits in my notes next to it, because the versions differ a lot: $0.045 per song on one, $0.15 on the next one up. An alias that silently moves would break both the cost math and the sound. The other reason I pin is that models get retired out from under you: Suno retired a whole run of its models on September 9, out from under every unofficial wrapper built on them. A name I chose, in a file I can read, at least means a change has to go through me.