OpenAI model pricing detail
This page answers the GPT-5.4 pricing query directly. It keeps the current short-context, long-context, and batch rows in one view, then shows the fit and hosted-tool signals that decide whether the flagship row is actually worth paying for.
Price rows
The pricing page lists three live GPT-5.4 rows that matter operationally: short context, long context, and short-context batch.
| Mode | Context | Input | Cached input | Output | Decision read |
|---|---|---|---|---|---|
| Standard | Short | $2.50 per 1M tokens | $0.25 per 1M tokens | $15 per 1M tokens | This is the lowest live GPT-5.4 row for direct API usage before hosted tools are added. |
| Standard | Long | $5.00 per 1M tokens | $0.50 per 1M tokens | $23 per 1M tokens | The long-context row is materially more expensive and should be priced separately from short-context use. |
| Batch | Short only | $1.25 per 1M tokens | $0.13 per 1M tokens | $7.50 per 1M tokens | Batch only applies to the short-context row. There is no separate long-context batch row listed today. |
Fit signals
Context size and tool support are what usually keep GPT-5.4 in the path after the initial price comparison starts.
The current OpenAI limits summary lists GPT-5.4 at 1,048,576 tokens. If the workload actually uses that headroom, the row is competing on fit, not just on token price.
Sources
OpenAI currently lists GPT-5.4 with Functions, web search, file search, skills, image generation, code interpreter, MCP. If the workflow depends on that breadth, a cheaper model row is not a full replacement.
Sources
GPT-5.4 short and GPT-5.4 long are not interchangeable estimates. If the workload lands on the long-context row, the flagship path needs its own explicit price read before a cheaper alternative is considered.
Sources
Workload snapshots
These snapshots keep the arithmetic simple so the row can be understood before a broader calculator is opened.
When this row is misleading
These are the three places where the flagship row often gets over-trusted in a budget discussion.
File search storage bills at $0.10 per GB per day after the first 1 free GB, and file-search calls bill at $2.50 per 1K calls. That often matters more than shaving a few dollars off the token row.
Sources
If the workflow uses OpenAI web search, the estimate depends on the search path and its separate call pricing, not just on GPT-5.4 input and output tokens.
Sources
If code interpreter or runtime-heavy workflows stay in the path, the runtime line remains a separate cost surface rather than a hidden part of the GPT-5.4 token row.
Official sources
This page keeps the source set narrow so a pricing row can stay auditable instead of drifting into summary-only advice.
Source of record for GPT-5.4 short, long, and batch token pricing.
Source of record for the published context window, max output tokens, and built-in tool coverage that change the value of the pricing row.
Continue the site
Use the groups below to move laterally through the decision, not back out into another doc hunt.
Related pages
Stay in the same decision neighborhood instead of backing out to search.
Model pricing, hosted-tool costs, and fit constraints that materially change the operating estimate.
Open pageSingle-model pricing brief for GPT-5 mini across standard and batch rows.
Open pageLimits brief for GPT-5.4 versus GPT-5 mini context windows, output caps, and tool support.
Open pageCompare pages
Open the pages that turn this topic into a side-by-side decision.
Replacement pages
Use the likely substitutes, migration targets, or fallback choices as the next click.
Single-model pricing brief for GPT-5 mini across standard and batch rows.
Open pageInteractive calculator for model tokens, hosted tools, and runtime in one estimate.
Open pageTool-cost brief for file search pricing across storage, tool calls, and model-token exposure.
Open pageSource category pages
Trace the source families behind this page instead of opening random docs in isolation.