OpenAI: GPT-5.4 Nano (batch)
GPT-5.4 Nano (batch) is an OpenAI model that accepts text, image, and file inputs and returns text. It offers a 400,000-token context window with up to 128,000 tokens of output per response, which leaves room for long documents or large codebases in a single request. It supports tool calling and reasoning; the available facts do not confirm structured output support. At $0.10 per million input tokens and $0.625 per million output tokens, it belongs on the shortlist for high-volume work where cost matters more than peak capability. Its blended score of 68.1 rests on only three benchmarks, so treat its standing as provisional. Within that set, results are uneven: 92.5 on the coding index, 63.1 on the general index, and 45.4 on the agentic index. That profile favors code generation and review over multi-step agent workflows, where buyers should test it carefully before committing.
Benchmark results
Independent, published benchmarks. Blended score 68.1 across 3 benchmarks, last refreshed 2026-11-08. How scoring works →
| Benchmark | Measures | Score |
|---|---|---|
| AI Index | broad capability composite | 63.1 |
| AI Index Coding | software engineering tasks | 92.5 |
| AI Index Agentic | multi-step tool-using tasks | 45.4 |
- Model ID
- openai/gpt-5.4-nano:batch
- Vendor
- openai
- Released
- March 2026
- Tokenizer
- GPT
- Input Modalities
- file, image, text
- Output Modalities
- text
- Max Output
- 128,000 tokens
- Tool Calling
- ✓ supported
- Structured Output
- ✓ supported
- Reasoning Mode
- ✓ supported
- Vision
- ✓ accepts images
- Audio
- no
- Moderated
- yes
What it costs in practice
Computed from the current $0.10/M input and $0.62/M output rates. Run your own numbers →
| Job | Tokens | Cost |
|---|---|---|
| Summarize a 50-page report | 30k in / 1.5k out | under $0.01 |
| Classify 1,000 customer emails | 500k in / 50k out | $0.08 |
| A month of a busy support chatbot | 5M in / 2M out | $1.75 |
Strong choice for
Category rankings
Where OpenAI: GPT-5.4 Nano (batch) places across the 15 categories it ranks in. How we rank →
| # | Category | Score |
|---|---|---|
| #5 | Bulk Data LabelingData · of 25 ranked | 134 |
| #7 | Dataset AnnotationResearch · of 25 ranked | 143 |
| #8 | Customer SupportBusiness · of 25 ranked | 133 |
| #10 | Short-Form SummarizationWriting · of 25 ranked | 130 |
| #10 | Email DraftingWriting · of 25 ranked | 126 |
| #10 | Language LearningEducation · of 25 ranked | 126 |
| #10 | Chat CompanionPersonal · of 25 ranked | 130 |
| #11 | JSON ExtractionData · of 25 ranked | 147 |
| #13 | Code CompletionCode · of 25 ranked | 133 |
| #14 | Trivia & General KnowledgePersonal · of 25 ranked | 120 |
| #15 | Image CaptioningVision · of 25 ranked | 120 |
| #24 | Sales / Cold EmailBusiness · of 25 ranked | 122 |
| #24 | Job Application DraftingBusiness · of 25 ranked | 122 |
| #24 | Journaling HelperPersonal · of 25 ranked | 122 |
| #24 | Recipe GenerationPersonal · of 25 ranked | 122 |
Similar models
OpenAI: GPT-6.1 Terra Pro
OpenAI: GPT-6.1 Terra Pro (batch)
OpenAI: GPT-6.1 Terra
OpenAI: GPT-6.1 Terra (batch)
OpenAI: GPT-6.1 Sol Pro
OpenAI: GPT-6.1 Sol Pro (batch)
Quick answers
- How much does OpenAI: GPT-5.4 Nano (batch) cost?
- $0.10 per million input tokens and $0.62 per million output tokens.
- What is OpenAI: GPT-5.4 Nano (batch)'s context window?
- 400,000 tokens, roughly 600 pages of text in a single request.
- Does OpenAI: GPT-5.4 Nano (batch) support tool calling?
- It supports tool calling, structured output, a reasoning mode.
- Can OpenAI: GPT-5.4 Nano (batch) process images?
- Yes, it accepts image input.