Standard input and output
Input is the prompt and context sent to the model; output is generated text. Match the model ID, displayed mode, and context tier before comparing rates.
View current model pricesOpenAI API Pricing
Review current OpenAI API prices for GPT models, including input, cached input, and output token rates per 1M tokens, verification dates, and official OpenAI source links.
- MODELS COVERED
- 19
- DATA SOURCE
- Official OpenAI
- PRICING UNIT
- Multiple billing units
- LAST VERIFIED
- Oct 9, 2026
Pricing modes at a glance
How to read OpenAI API pricing
Compare source-linked model rates in USD per 1M tokens, inspect context tiers, and reuse the same reference in the model pricing comparison or the machine-readable pricing dataset. The table gives unit prices; the calculator estimates your workload cost.
Latest model verification: Oct 9, 2026. This is the latest date across model records, not a claim that every price was checked that day. Read each price record's Verified, Checked, and official source links beside the displayed prices. Missing rates stay Not listed; they do not mean free usage. Review-required or legacy fallback rows are labeled and must not be treated as verified current prices.
Cached input
Cached Input is the rate for eligible cache reads, not the ordinary Input rate. Repeated content is not automatically a cache hit. Cache writes can have a separate rate; use the matching model, mode, and context records in the dataset rather than counting the same tokens twice.
Compare input and cached inputBatch pricing
Batch is a processing mode for eligible asynchronous requests, not a separate model. Use listed Batch Input and Batch Output rates for the matching model and context. Not listed does not establish availability or a discount; check the official source for supported workloads and completion windows.
Compare Standard and Batch workload costLong-context and tiered pricing
Some models change rates when a token threshold is crossed. Each model's tier table specifies the threshold, token basis, and whole-request rule; there is no provider-wide threshold.
Read model-specific context tiersNon-token components
Minute-based and other native-unit charges are separate from text-token rates. The records below retain their billing units. This page does not establish complete image, audio, or tool coverage; verify any additional charges with the official source.
Illustrative single-request cost
GPT-5.4 mini (gpt-5.4-mini): 800 input tokens × $0.750000 / 1,000,000 + 350 output tokens × $4.50 / 1,000,000 = $0.002175 per request. No cache hits or writes, Batch discount, tool charges, retries, or regional adjustments are included. This is a Standard text-token illustration, not a bill or benchmark.
Check this example's model rates and evidenceTo apply your own token counts and request volume, calculate your OpenAI API workload cost.
Non-token billing
Non-token prices retain their native billing units.
These records are published separately from token pricing and are excluded from token calculators and the model token-price comparison.
GPT-Live 1
- API price
- $0.05 / minute
- Billing dimension
- Session duration
- Model ID
- gpt-live-1
- Provider
- OpenAI
Evidence
Dedicated model pages
Pricing by model
Open a model page for its verified source, workload examples, price history, lifecycle details, and nearby alternatives.
OpenAI API pricing table
OpenAI API pricing table
Compare OpenAI input token, output token, and cached input pricing from the existing source-linked dataset. Not listed values are never treated as zero.
| Model | Input | Output | Cached Input | Batch Input | Batch Output | Unit | Status |
|---|---|---|---|---|---|---|---|
chat-latestModel ID: chatgpt-chat-latestProvider: OpenAIDisplayed pricing: StandardEvidence | Input · USD per 1M input tokens$5.00 | Output · USD per 1M output tokens$30.00 | Cached input · USD per 1M cached input tokens$0.500 | Batch input · USD per 1M input tokensNot listed | Batch output · USD per 1M output tokensNot listed | USD per 1M tokens | specializedChatGPT category |
GPT-4.1Model ID: gpt-4.1Provider: OpenAIEvidence | Input · USD per 1M input tokens$2.00 | Output · USD per 1M output tokens$8.00 | Cached input · USD per 1M cached input tokensNot listed | Batch input · USD per 1M input tokensNot listed | Batch output · USD per 1M output tokensNot listed | USD per 1M tokens | legacyLegacy |
GPT-4.1 miniModel ID: gpt-4.1-miniProvider: OpenAIEvidence | Input · USD per 1M input tokens$0.400 | Output · USD per 1M output tokens$1.60 | Cached input · USD per 1M cached input tokensNot listed | Batch input · USD per 1M input tokensNot listed | Batch output · USD per 1M output tokensNot listed | USD per 1M tokens | legacyLegacy |
GPT-4.1 nanoModel ID: gpt-4.1-nanoProvider: OpenAIEvidence | Input · USD per 1M input tokens$0.100 | Output · USD per 1M output tokens$0.400 | Cached input · USD per 1M cached input tokensNot listed | Batch input · USD per 1M input tokensNot listed | Batch output · USD per 1M output tokensNot listed | USD per 1M tokens | legacyLegacy |
GPT-5.3 CodexModel ID: gpt-5.3-codexProvider: OpenAIEvidence | Input · USD per 1M input tokens$1.75 | Output · USD per 1M output tokens$14.00 | Cached input · USD per 1M cached input tokens$0.175 | Batch input · USD per 1M input tokensNot listed | Batch output · USD per 1M output tokensNot listed | USD per 1M tokens | specializedCodex |
GPT-5.4Model ID: gpt-5.4Provider: OpenAIDisplayed pricing: StandardEvidence | Input · USD per 1M input tokens$2.50 | Output · USD per 1M output tokens$15.00 | Cached input · USD per 1M cached input tokens$0.250 | Batch input · USD per 1M input tokens$1.25 | Batch output · USD per 1M output tokens$7.50 | USD per 1M tokens | latestStandard |
GPT-5.4 miniModel ID: gpt-5.4-miniProvider: OpenAIDisplayed pricing: StandardEvidence | Input · USD per 1M input tokens$0.750 | Output · USD per 1M output tokens$4.50 | Cached input · USD per 1M cached input tokens$0.075 | Batch input · USD per 1M input tokens$0.375 | Batch output · USD per 1M output tokens$2.25 | USD per 1M tokens | latestStandard |
GPT-5.4 nanoModel ID: gpt-5.4-nanoProvider: OpenAIEvidence Verified: Not availableChecked: Not availableOfficial sources (3): Official source 1Official source 2Official source 3 | Input · USD per 1M input tokens$0.200 | Output · USD per 1M output tokens$1.25 | Cached input · USD per 1M cached input tokens$0.020 | Batch input · USD per 1M input tokensNot listed | Batch output · USD per 1M output tokensNot listed | USD per 1M tokens | latestStandard |
GPT-5.4 ProModel ID: gpt-5.4-proProvider: OpenAIDisplayed pricing: StandardEvidence | Input · USD per 1M input tokens$30.00 | Output · USD per 1M output tokens$180.00 | Cached input · USD per 1M cached input tokensNot listed | Batch input · USD per 1M input tokens$15.00 | Batch output · USD per 1M output tokens$90.00 | USD per 1M tokens | latestPremium |
GPT-5.5Model ID: gpt-5.5Provider: OpenAIDisplayed pricing: Standard · Short context · ≤272k prompt tokens · whole requestEvidence | Input · USD per 1M input tokens$5.00 | Output · USD per 1M output tokens$30.00 | Cached input · USD per 1M cached input tokens$0.500 | Batch input · USD per 1M input tokens$2.50 | Batch output · USD per 1M output tokens$15.00 | USD per 1M tokens | latestStandard |
GPT-5.5 ProModel ID: gpt-5.5-proProvider: OpenAIDisplayed pricing: StandardEvidence | Input · USD per 1M input tokens$30.00 | Output · USD per 1M output tokens$180.00 | Cached input · USD per 1M cached input tokensNot listed | Batch input · USD per 1M input tokens$15.00 | Batch output · USD per 1M output tokens$90.00 | USD per 1M tokens | latestPremium |
GPT-5.6 LunaModel ID: gpt-5.6-lunaProvider: OpenAIDisplayed pricing: Standard · Short context · ≤272k prompt tokens · whole requestEvidence | Input · USD per 1M input tokens$0.200 | Output · USD per 1M output tokens$1.20 | Cached input · USD per 1M cached input tokens$0.020 | Batch input · USD per 1M input tokens$0.100 | Batch output · USD per 1M output tokens$0.600 | USD per 1M tokens | latestStandard |
GPT-5.6 SolModel ID: gpt-5.6-solProvider: OpenAIDisplayed pricing: Standard · Short context · ≤272k prompt tokens · whole requestEvidence | Input · USD per 1M input tokens$4.00 | Output · USD per 1M output tokens$20.00 | Cached input · USD per 1M cached input tokens$0.400 | Batch input · USD per 1M input tokens$2.00 | Batch output · USD per 1M output tokens$10.00 | USD per 1M tokens | latestStandard |
GPT-5.6 TerraModel ID: gpt-5.6-terraProvider: OpenAIDisplayed pricing: Standard · Short context · ≤272k prompt tokens · whole requestEvidence | Input · USD per 1M input tokens$2.00 | Output · USD per 1M output tokens$12.00 | Cached input · USD per 1M cached input tokens$0.200 | Batch input · USD per 1M input tokens$1.00 | Batch output · USD per 1M output tokens$6.00 | USD per 1M tokens | latestStandard |
GPT-6 AstraModel ID: gpt-6-astraProvider: OpenAIDisplayed pricing: Standard · Short context · ≤272k input tokens · whole requestEvidence | Input · USD per 1M input tokens$10.00 | Output · USD per 1M output tokens$50.00 | Cached input · USD per 1M cached input tokens$1.00 | Batch input · USD per 1M input tokens$5.00 | Batch output · USD per 1M output tokens$25.00 | USD per 1M tokens | latestactive |
GPT-6 LunaModel ID: gpt-6-lunaProvider: OpenAIDisplayed pricing: Standard · Short context · ≤272k input tokens · whole requestEvidence | Input · USD per 1M input tokens$0.100 | Output · USD per 1M output tokens$0.500 | Cached input · USD per 1M cached input tokens$0.010 | Batch input · USD per 1M input tokens$0.050 | Batch output · USD per 1M output tokens$0.250 | USD per 1M tokens | latestactive |
GPT-6 SolModel ID: gpt-6-solProvider: OpenAIDisplayed pricing: Standard · Short context · ≤272k input tokens · whole requestEvidence | Input · USD per 1M input tokens$2.00 | Output · USD per 1M output tokens$10.00 | Cached input · USD per 1M cached input tokens$0.200 | Batch input · USD per 1M input tokens$1.00 | Batch output · USD per 1M output tokens$5.00 | USD per 1M tokens | latestactive |
GPT-6.1 SolModel ID: gpt-6.1-solProvider: OpenAIDisplayed pricing: Standard · Short context · ≤272k input tokens · whole requestEvidence | Input · USD per 1M input tokens$2.00 | Output · USD per 1M output tokens$10.00 | Cached input · USD per 1M cached input tokens$0.100 | Batch input · USD per 1M input tokens$1.00 | Batch output · USD per 1M output tokens$5.00 | USD per 1M tokens | latestactive |
Use these OpenAI pricing rows as planning inputs: calculate your monthly API cost with realistic token counts, or compare this provider with other AI models before routing production traffic.
For deeper pricing context, read the monthly OpenAI API pricing change history and GPT-6 and Claude Opus 5.5 API costs and how the OpenAI Free usage tier and low-cost paid path work.
Canonical pricing tiers
GPT-5.5 short- and long-context pricing.
These standard API prices come from the canonical tier records attached to this model. Values are shown in USD per 1M tokens and are not inferred from the base row.
| Context tier | Total prompt token rule | Regular input | Cached input | Output | Unit |
|---|---|---|---|---|---|
| Short contextshort | ≤272k prompt tokens | $5.00 | $0.500 | $30.00 | USD / 1M tokens |
| Long contextlong | >272k prompt tokens | $10.00 | $1.00 | $45.00 | USD / 1M tokens |
Total prompt tokens
Use this model's token rule in the tier table above. Cached input is included in the threshold count; the threshold and token basis belong to this model, not every OpenAI model.
Whole-request pricing
When a tier's threshold condition matches, that tier prices the entire standard request. This is a tier switch, not progressive pricing across token bands.
Price-record evidence
GPT-5.5
Model ID: gpt-5.5
Provider: OpenAI
Evidence
Canonical pricing tiers
GPT-5.6 Luna short- and long-context pricing.
These standard API prices come from the canonical tier records attached to this model. Values are shown in USD per 1M tokens and are not inferred from the base row.
| Context tier | Total prompt token rule | Regular input | Cached input | Output | Unit |
|---|---|---|---|---|---|
| Short contextshort | ≤272k prompt tokens | $0.200 | $0.020 | $1.20 | USD / 1M tokens |
| Long contextlong | >272k prompt tokens | $0.400 | $0.040 | $1.80 | USD / 1M tokens |
Total prompt tokens
Use this model's token rule in the tier table above. Cached input is included in the threshold count; the threshold and token basis belong to this model, not every OpenAI model.
Whole-request pricing
When a tier's threshold condition matches, that tier prices the entire standard request. This is a tier switch, not progressive pricing across token bands.
Price-record evidence
GPT-5.6 Luna
Model ID: gpt-5.6-luna
Provider: OpenAI
Evidence
Canonical pricing tiers
GPT-5.6 Sol short- and long-context pricing.
These standard API prices come from the canonical tier records attached to this model. Values are shown in USD per 1M tokens and are not inferred from the base row.
| Context tier | Total prompt token rule | Regular input | Cached input | Output | Unit |
|---|---|---|---|---|---|
| Short contextshort | ≤272k prompt tokens | $4.00 | $0.400 | $20.00 | USD / 1M tokens |
| Long contextlong | >272k prompt tokens | $8.00 | $0.800 | $30.00 | USD / 1M tokens |
Total prompt tokens
Use this model's token rule in the tier table above. Cached input is included in the threshold count; the threshold and token basis belong to this model, not every OpenAI model.
Whole-request pricing
When a tier's threshold condition matches, that tier prices the entire standard request. This is a tier switch, not progressive pricing across token bands.
Price-record evidence
GPT-5.6 Sol
Model ID: gpt-5.6-sol
Provider: OpenAI
Evidence
Canonical pricing tiers
GPT-5.6 Terra short- and long-context pricing.
These standard API prices come from the canonical tier records attached to this model. Values are shown in USD per 1M tokens and are not inferred from the base row.
| Context tier | Total prompt token rule | Regular input | Cached input | Output | Unit |
|---|---|---|---|---|---|
| Short contextshort | ≤272k prompt tokens | $2.00 | $0.200 | $12.00 | USD / 1M tokens |
| Long contextlong | >272k prompt tokens | $4.00 | $0.400 | $18.00 | USD / 1M tokens |
Total prompt tokens
Use this model's token rule in the tier table above. Cached input is included in the threshold count; the threshold and token basis belong to this model, not every OpenAI model.
Whole-request pricing
When a tier's threshold condition matches, that tier prices the entire standard request. This is a tier switch, not progressive pricing across token bands.
Price-record evidence
GPT-5.6 Terra
Model ID: gpt-5.6-terra
Provider: OpenAI
Evidence
Canonical pricing tiers
GPT-6 Astra short- and long-context pricing.
These standard API prices come from the canonical tier records attached to this model. Values are shown in USD per 1M tokens and are not inferred from the base row.
| Context tier | Input token rule | Regular input | Cached input | Output | Unit |
|---|---|---|---|---|---|
| Short contextshort | ≤272k input tokens | $10.00 | $1.00 | $50.00 | USD / 1M tokens |
| Long contextlong | >272k input tokens | $20.00 | $2.00 | $75.00 | USD / 1M tokens |
Input token threshold
Use this model's token rule in the tier table above. Cached input is included in the threshold count; the threshold and token basis belong to this model, not every OpenAI model.
Whole-request pricing
When a tier's threshold condition matches, that tier prices the entire standard request. This is a tier switch, not progressive pricing across token bands.
Price-record evidence
GPT-6 Astra
Model ID: gpt-6-astra
Provider: OpenAI
Evidence
Canonical pricing tiers
GPT-6 Luna short- and long-context pricing.
These standard API prices come from the canonical tier records attached to this model. Values are shown in USD per 1M tokens and are not inferred from the base row.
| Context tier | Input token rule | Regular input | Cached input | Output | Unit |
|---|---|---|---|---|---|
| Short contextshort | ≤272k input tokens | $0.100 | $0.010 | $0.500 | USD / 1M tokens |
| Long contextlong | >272k input tokens | $0.200 | $0.020 | $0.750 | USD / 1M tokens |
Input token threshold
Use this model's token rule in the tier table above. Cached input is included in the threshold count; the threshold and token basis belong to this model, not every OpenAI model.
Whole-request pricing
When a tier's threshold condition matches, that tier prices the entire standard request. This is a tier switch, not progressive pricing across token bands.
Price-record evidence
GPT-6 Luna
Model ID: gpt-6-luna
Provider: OpenAI
Evidence
Canonical pricing tiers
GPT-6 Sol short- and long-context pricing.
These standard API prices come from the canonical tier records attached to this model. Values are shown in USD per 1M tokens and are not inferred from the base row.
| Context tier | Input token rule | Regular input | Cached input | Output | Unit |
|---|---|---|---|---|---|
| Short contextshort | ≤272k input tokens | $2.00 | $0.200 | $10.00 | USD / 1M tokens |
| Long contextlong | >272k input tokens | $4.00 | $0.400 | $15.00 | USD / 1M tokens |
Input token threshold
Use this model's token rule in the tier table above. Cached input is included in the threshold count; the threshold and token basis belong to this model, not every OpenAI model.
Whole-request pricing
When a tier's threshold condition matches, that tier prices the entire standard request. This is a tier switch, not progressive pricing across token bands.
Price-record evidence
GPT-6 Sol
Model ID: gpt-6-sol
Provider: OpenAI
Evidence
Canonical pricing tiers
GPT-6.1 Sol short- and long-context pricing.
These standard API prices come from the canonical tier records attached to this model. Values are shown in USD per 1M tokens and are not inferred from the base row.
| Context tier | Input token rule | Regular input | Cached input | Output | Unit |
|---|---|---|---|---|---|
| Short contextshort | ≤272k input tokens | $2.00 | $0.100 | $10.00 | USD / 1M tokens |
| Long contextlong | >272k input tokens | $4.00 | $0.200 | $15.00 | USD / 1M tokens |
Input token threshold
Use this model's token rule in the tier table above. Cached input is included in the threshold count; the threshold and token basis belong to this model, not every OpenAI model.
Whole-request pricing
When a tier's threshold condition matches, that tier prices the entire standard request. This is a tier switch, not progressive pricing across token bands.
Price-record evidence
GPT-6.1 Sol
Model ID: gpt-6.1-sol
Provider: OpenAI
Evidence
OpenAI API pricing and usage planning
Turn token assumptions into a monthly API cost estimate.
Use these planning modules with the calculator instead of treating the pricing table as a final bill. Your actual OpenAI API pricing depends on the request shape, model route, cache behavior, and production traffic.
OpenAI API pricing table
Review current OpenAI model rows with input, output, cached input, unit, status, and official source links. The table below stays connected to the existing project pricing data instead of using hand-written prices.
Estimate your monthly OpenAI API cost
Move realistic input tokens, output tokens, requests per month, and model assumptions into the calculator to estimate monthly OpenAI API cost for an app, chatbot, workflow, or internal assistant.
Open the AI API cost calculatorCompare Standard and Batch API cost
Upload CSV or XLSX workloads, select the text column, include expected output tokens, and compare eligible OpenAI Standard and Batch rates without changing the pricing reference on this page.
OpenAI Batch API cost calculatorOpenAI Responses API pricing
For Responses API workloads, estimate the same billing drivers: model choice, input tokens, output tokens, cached input eligibility, tool context, and repeated system instructions. Always verify final rules with official OpenAI billing pages.
Customer support chatbot workload
Estimate a complete support conversation from system instructions, retrieved help-center context, chat history, user messages, assistant replies, and regenerations rather than treating it as one request.
See a full customer support chatbot API cost exampleOpenAI API monthly cost examples
Common GPT API workloads with calculated monthly examples.
These examples reuse the same provider pricing rows and token-cost helper as the table. Estimates use the selected API model and current listed token prices; actual costs depend on workload, caching, retries, and provider billing rules.
Customer support chatbot
A support chatbot that answers product questions with concise responses and repeated help-center context.
- Input
- 800 tokens
- Output
- 350 tokens
- Estimated monthly cost
- $108.75
- Model row
- GPT-5.4 mini (gpt-5.4-mini)
Article summarizer
A summarizer that sends long articles or transcripts and returns shorter editorial briefs.
- Input
- 3,000 tokens
- Output
- 500 tokens
- Estimated monthly cost
- $27.00
- Model row
- GPT-5.4 mini (gpt-5.4-mini)
SaaS AI assistant
An app assistant that combines user messages, account context, tool results, and generated replies.
- Input
- 1,500 tokens
- Output
- 700 tokens
- Estimated monthly cost
- $513.00
- Model row
- GPT-5.4 mini (gpt-5.4-mini)
Cheapest OpenAI API models
Find lower-cost GPT routes without ignoring quality.
The lowest token price is a starting point, not a complete decision. For production, compare total workflow cost, answer quality, output length, cache hit rate, retries, and fallback behavior.
GPT-4.1 nano
- Input
- $0.100
- Output
- $0.400
- Cached input
- Not listed
GPT-6 Luna
- Input
- $0.100
- Output
- $0.500
- Cached input
- $0.010
GPT-5.6 Luna
- Input
- $0.200
- Output
- $1.20
- Cached input
- $0.020
Cheapest model cards are derived from the same pricing rows shown on this page and are informational only. Verify current OpenAI API pricing, billing rules, and invoices before final purchasing decisions.
Cost examples
- 1M input tokens
- $5.00
- Model
- chat-latest
- 1M output tokens
- $30.00
- Model
- chat-latest
- 10M input + 2M output
- $110.00
- Model
- chat-latest
Calculate OpenAI API costs with your own usage assumptions.
Use the calculator for monthly app usage, ChatGPT API cost examples, cached input scenarios, and model-routing comparisons.
Pricing data is informational and may lag provider changes. Always verify final billing decisions against provider pricing pages, contracts, dashboards, and invoices.
OpenAI API pricing FAQ
How much does the OpenAI API cost?
OpenAI API cost depends on the selected GPT model, input tokens, cached input eligibility, output tokens, request volume, retries, Batch API usage, and monthly traffic. Use the pricing table and calculator together to estimate your own workload instead of relying on a single average price.
How is OpenAI API pricing calculated per 1M tokens?
Text-token rates use a per-1M-token unit. For the matching model, mode, and context tier, multiply uncached input, eligible cache reads, any separately priced cache writes, and output by their applicable rates, then divide by one million. Do not count the same input tokens in multiple categories. Add applicable non-token charges separately.
What are input, cached input, and output tokens?
Input is the prompt and context sent to the model. Cached Input is the rate for eligible cache reads, not ordinary input or cache writes. Repeated content must qualify under the model and processing-mode rules. Output is generated text and is priced separately; use official documentation to interpret billed token categories.
Is an API key charged separately from API usage?
An API key is an access credential, not a separate product price on this page. OpenAI API billing is generally based on actual usage, account terms, credits, limits, and invoices shown in the official OpenAI billing dashboard.
Is ChatGPT pricing the same as API pricing?
No. ChatGPT subscriptions and OpenAI API pricing are separate billing surfaces. A ChatGPT plan does not automatically make API usage free, and API usage should be estimated from model token prices, requests, and your OpenAI account billing terms.
How can monthly OpenAI API cost be estimated?
Estimate one representative request, then multiply by requests per user, active users, days per month, retries, and fallback calls. Use the API cost calculator when you need to model monthly GPT API spend from realistic input and output token assumptions.
Does OpenAI offer Batch API pricing?
Batch is a processing mode for eligible asynchronous workloads, not a separate model. Compare the listed Batch Input and Batch Output rates with Standard rates for the same model and context. Missing rates do not establish availability or free usage. Check the official source for model support, completion windows, and billing rules.
Do long-context requests use different prices?
Some models have context-based pricing tiers. Read the model-specific token threshold, token basis, and whole-request rule shown in its tier table. Do not apply one model's threshold to every OpenAI model or treat a whole-request tier switch as progressive billing. Models without listed tiers do not acquire a tier rule from another row.
Where can users verify official OpenAI pricing?
Use the external official OpenAI pricing page, OpenAI API documentation, your OpenAI billing dashboard, contracts, and invoices as the final source of truth. AICostBudget keeps source links visible, but it is not an official OpenAI pricing page.