OpenAI API Pricing 2026:
GPT-5.4, GPT-5.4 mini, GPT-5.4 nano & Current Models
Current OpenAI API pricing for the GPT-5.4 family, plus GPT-4o mini for comparison. Includes Batch API discounts, web search tool pricing, and model selection guidance. Last verified: 2026-04-01.
GPT-5.4 Family — Current Production Models
| Model | Input / 1M tokens | Output / 1M tokens | Context | Best for |
|---|---|---|---|---|
| GPT-5.4 | $2.50 | $15.00 | 1M tokens | Complex reasoning, coding, hard tasks. Standard pricing under 270k context. |
| GPT-5.4 mini | $0.75 | $4.50 | 128K tokens | Balanced cost/quality at scale |
| GPT-5.4 nano | $0.20 | $1.25 | 128K tokens | Ultra-high volume, simple tasks |
Older / Comparison Models
These models remain available via the API and are still common in existing integrations. For new projects, the GPT-5.4 family is the recommended starting point.
| Model | Input / 1M | Output / 1M | Context | Notes |
|---|---|---|---|---|
| GPT-5 mini | $0.25 | $2.00 | 128K | Legacy bridge model |
| GPT-5 nano | $0.05 | $0.40 | 128K | Legacy low-cost option |
| GPT-4o | $2.50 | $10.00 | 128K | Comparison only — older general model |
| GPT-4o mini | $0.15 | $0.60 | 128K | Strong budget model — still widely used |
Previous Reasoning Models
o3 and o4-mini have been succeeded by the GPT-5 family. They remain on the API for existing integrations but should not be used as the default for new projects.
OpenAI Batch API — 50% Discount for Non-Real-Time Work
OpenAI's Batch API offers a 50% discount on all standard model pricing for asynchronous workloads (evals, data processing, classification). Batches complete within 24 hours.
| Model | Standard Input | Batch Input | Standard Output | Batch Output |
|---|---|---|---|---|
| GPT-5.4 | $2.50 | $1.25 | $15.00 | $7.50 |
| GPT-5.4 mini | $0.75 | $0.375 | $4.50 | $2.25 |
| GPT-5.4 nano | $0.20 | $0.10 | $1.25 | $0.625 |
Which OpenAI Model Is Best for Cost-Sensitive Apps?
| Use case | Recommended model | Reason |
|---|---|---|
| Complex reasoning, research, coding | GPT-5.4 | Flagship model, 1M context, top quality |
| Balanced production workloads | GPT-5.4 mini | Good quality at $0.75/M — efficient middle tier |
| High-volume simple tasks | GPT-5.4 nano | Cheapest 5.4-class model at $0.20/M |
| Legacy budget integrations | GPT-4o mini | $0.15/M — still the absolute cheapest OpenAI option |
| Async classification / evals | Batch API (any model) | 50% off all models, 24h turnaround |
When GPT-4o mini Is Still the Best Budget Choice
GPT-4o mini at $0.15/M input is still cheaper than GPT-5.4 nano ($0.20/M). If you have high-volume, simple classification or chat tasks already running on GPT-4o mini, migrating purely for cost savings may not be compelling. GPT-5.4 nano wins on quality-to-cost ratio if you need better instruction-following or more nuanced output.
Model Selection by Use Case
- Customer support (volume, simple): GPT-5.4 nano or GPT-4o mini
- Customer support (complex, nuanced): GPT-5.4 mini
- Code generation / review: GPT-5.4
- Document summarization (long): GPT-5.4 with 1M context
- Content generation at scale: GPT-5.4 mini + Batch API
- Classification pipeline: GPT-5.4 nano + Batch API
OpenAI API Access
- Sign up at platform.openai.com — immediate access, no approval needed
- Free credits for new accounts ($5–$18 depending on current promotion)
- Pay-as-you-go after free tier — no monthly minimum
- Enterprise: custom pricing with dedicated capacity — contact OpenAI sales
Frequently Asked Questions
What is the difference between OpenAI API and ChatGPT?
The OpenAI API is the developer-facing service where you pay per token based on your application usage. ChatGPT is the consumer web and mobile product with flat-rate subscriptions ($20/month for Plus, $200/month for Pro). Building a product requires the API, not a ChatGPT subscription.
Is GPT-5.4 cheaper than Claude Sonnet 4.6?
GPT-5.4 ($2.50/M input) is slightly cheaper than Claude Sonnet 4.6 ($3.00/M input) on a per-token basis. Both support 1M context at standard rates. Claude offers more aggressive prompt caching discounts (90% off reads), which can make it cheaper for workloads with large repeated system prompts.
How does the OpenAI Batch API save money?
The Batch API gives a 50% discount on all standard models for jobs submitted as batches (up to 10,000 requests). Jobs complete within 24 hours. It's ideal for non-latency-sensitive tasks like evals, bulk classification, content processing, and dataset enrichment.
Does OpenAI charge extra for the 1M context window?
GPT-5.4 supports up to 1M tokens in a single context at standard pricing, but sessions exceeding approximately 270k tokens may apply a higher rate. For most applications, standard pricing applies. Check OpenAI's official pricing page for current long-context thresholds.
Calculate Your OpenAI API Costs
Compare GPT-5.4 vs Claude vs Gemini for your specific usage volume.
Open API Cost Calculator