I'm trying to find the current pricing for the GPT-5.6 model family in Azure Foundry, but I can't locate any entries in the public retail pricing API or the official Azure OpenAI pricing page. The model monitor is currently showing an estimated total cost of $0. Is pricing simply not published yet, or is there another source for the actual rates?
2 Answers
Be careful with the $0 estimate. One team reported receiving a roughly $5,000 bill from a small number of unusually large requests, and their cost-management data was several times higher than their own usage tracking. They also saw charges for cache writes that weren’t exposed in the API response, making it difficult to pass those costs on to customers. I’d set strict spending limits and track token usage independently until the pricing is documented clearly.
It looks like the model may effectively be free or unpriced until Azure publishes official rates, but I wouldn’t rely on that assumption for production workloads. The monitoring estimate may just be missing the model’s pricing metadata rather than confirming that usage costs nothing.
That’s what I was wondering, but with the usage already reaching hundreds of millions of tokens, I’m hesitant to treat the zero estimate as reliable.