GPT-5.4 Mini
GPT-5.4 Mini targets frequent coding, computer interaction and delegated tasks. It has a smaller context window than the full GPT-5.4 model.
Total context and maximum input are separate
The 400,000-token window includes a maximum of 272,000 input tokens and up to 128,000 output tokens. Reasoning defaults to none. Responses supports tools such as shell, computer use and search; the model also supports Chat Completions and Batch.GPT-5.4 Mini official model documentation
Give delegated work a clear boundary
Practical guidance. For a specialist code task, provide the relevant files, expected behavior and the conditions for escalating uncertainty. Keep the task small enough to review as one change. Measure the accuracy of the handoff result before assigning many similar tasks.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | API |
| Model identifiers | gpt-5.4-minigpt-5.4-mini-2026-03-17 |
| Context window | 400,000 tokens Published model context; input and output share this budget GPT-5.4 Mini official model documentation |
| Maximum output | 128,000 tokens Published maximum output; bounded by the remaining context GPT-5.4 Mini official model documentation |
| Input price | 0.75 USD per million tokens Standard text processing at the reviewed date GPT-5.4 Mini official model documentation |
| Output price | 4.5 USD per million tokens Standard text processing at the reviewed date GPT-5.4 Mini official model documentation |
| Input and output | text, image in; text outGPT-5.4 Mini official model documentation |
| Maximum input | 272,000 tokensGPT-5.4 Mini official model documentation |
| API routes | Chat Completions (v1/chat/completions); Responses (v1/responses); Batch (v1/batch)GPT-5.4 Mini official model documentation |
Related models
GPT-5.4 is an available coding and professional-work model with a March 2026 snapshot.
Launch date not confirmedAPI- GPT-5.4 Pro
OpenAI · GPT-5.4
GPT-5.4 Pro is the higher-compute GPT-5.4 release for difficult problems.
Launch date not confirmedAPI