GPT-5.6 Sol Pricing Drops to $4/$10 per Million Tokens

GPT-5.6 Sol pricing has dropped to $4 per million input tokens and $10 per million output tokens, according to a post by @MatthewBerman. The post pairs those input/output rates with reported price reductions of 20% and 33%, respectively.

How the new rates affect token costs

The listed rates give developers a straightforward way to estimate usage costs while keeping input and output separate. One million output tokens costs 2.5 times as much as one million input tokens at the reported prices, so the overall cost impact will depend on how much a workload reads versus generates.

That distinction matters for applications with different token patterns. An input-heavy workflow will be influenced more by the $4 input rate, while a workflow that produces large volumes of text will be more affected by the $10 output rate. The announcement reports reductions for both categories, but the new prices do not make input and output tokens equally expensive.

A simple starting point for estimates

Using the reported rates, a basic token-cost estimate can be calculated as:

• Input cost: input tokens ÷ 1,000,000 × $4

• Output cost: output tokens ÷ 1,000,000 × $10

• Estimated total: input cost + output cost

For developers comparing usage costs, the immediate takeaway is the lower listed price per million tokens. Keeping the two calculations separate provides a more useful estimate than applying one blended rate, particularly because output tokens carry the higher reported price.