$$C_{\mathrm{total}} = C_{\mathrm{pre}} + C_{\mathrm{post}} + Q \cdot \bar{c}_{\mathrm{inf}}$$

Equation 1 · Foundation Models

OpenAI Model Systems from First Principles: Weights, Post-Training, and Inference Compute

The total compute bill has three parts: pretraining, post-training, and all the requests served afterwards.

Write the total lifetime computation of a deployed system as $C_{\mathrm{total}} = C_{\mathrm{pre}} + C_{\mathrm{post}} + Q \cdot \bar{c}_{\mathrm{inf}}$. where $C_{\mathrm{pre}}$ is pretraining compute, $C_{\mathrm{post}}$ is post-training compute, Q is the number of served requests over the system’s life, and $\bar{c}_{\mathrm{inf}}$ is the mean compute per request. Until roughly 2024 the third term was treated as approximately fixed for a given model, and public discussion of capability collapsed onto the first. That assumption no longer holds. On current OpenAI models, $\bar{c}_{\mathrm{inf}}$ is a parameter the caller sets: as verified on 8 August 2026, the model guidance documents a reasonin$g_e$ffort control taking the values none , low , medium , high , xhigh , and…

Meanings in this article

• $C_{\mathrm{total}}$: All computation spent on this deployed model system over its lifetime.
• $C_{\mathrm{pre}}$: The one-time computation used to learn from the initial training data.
• $C_{\mathrm{post}}$: The one-time computation used after pretraining to shape the model’s behavior.
• $Q$: How many requests the system serves during the period being counted.
• $\bar{c}_{\mathrm{inf}}$: The average computation used to answer one request. The bar means average.
• $Q \cdot \bar{c}_{\mathrm{inf}}$: Requests multiplied by average compute per request.

Equation guide → · Article →