Cloud deployment pricing
Azure LLM Pricing (July 2026)
Azure OpenAI GPT-5.5 Global Standard costs $5/$30 per million input/output tokens, with cached input at $0.50. The base Data Zone meter is $5.50/$33. Batch processing halves those rates to $2.50/$15 and $2.75/$16.50. Long-context GPT-5.5 traffic uses separate, higher meters.
Azure LLM pricing depends on more than the model name. The same GPT deployment can use Global Standard, Data Zone Standard, Regional Standard, Batch, Priority, or provisioned throughput, and each choice changes routing, residency, latency, or billing. The table below is a bounded USD snapshot of current pay-per-token meters, checked against Microsoft's public Azure OpenAI price sheet and Retail Prices API.
These are Microsoft retail list prices, not a quote. The selected Azure region, currency, contract, and deployment availability can change the bill. Confirm the final meter in the Azure calculator before committing production traffic.
Every Azure OpenAI API price per 1M tokens
| Model | Input $/M | Cached input $/M | Output $/M | Batch input $/M | Batch output $/M | Context | Effective | Price source |
|---|---|---|---|---|---|---|---|---|
| GPT-5.5 Global Standard | $5 | $0.5 | $30 | $2.5 | $15 | Standard meter | 2026-05-01 | Microsoft Azure |
| GPT-5.5 Data Zone Standard (base) | $5.5 | $0.55 | $33 | $2.75 | $16.5 | Standard meter | 2026-05-01 | Microsoft Azure |
| GPT-5.5 Global Standard, long context | $10 | $1 | $45 | $5 | $22.5 | Long-context meter | 2026-05-01 | Microsoft Azure |
| GPT-5.5 Data Zone, long context (base) | $11 | $1.1 | $49.5 | $5.5 | $24.75 | Long-context meter | 2026-05-01 | Microsoft Azure |
| GPT-5.4 Global Standard | $2.5 | $0.25 | $15 | $1.25 | $7.5 | <272K | 2026-03-01 | Microsoft Azure |
| GPT-5.4 Data Zone Standard (base) | $2.75 | $0.275 | $16.5 | $1.375 | $8.25 | <272K | 2026-03-01 | Microsoft Azure |
| GPT-5.4 Global Standard, long context | $5 | $0.5 | $22.5 | $2.5 | $11.25 | >272K | 2026-03-01 | Microsoft Azure |
| GPT-5.4 mini Global Standard | $0.75 | $0.075 | $4.5 | $0.375 | $2.25 | 1M | 2026-03-01 | Microsoft Azure |
| GPT-5.4 nano Global Standard | $0.2 | $0.02 | $1.25 | $0.1 | $0.625 | 1M | 2026-03-01 | Microsoft Azure |
USD retail list prices per 1M tokens. The Batch meters in Microsoft's retail feed are 50% below the corresponding standard rows; Microsoft also documents Global Batch as a 50% discount. Data Zone rows show the common base meter in the feed, while some APAC meters are higher. Regional, Priority, and provisioned-throughput rates are not represented.
Estimate your monthly Azure OpenAI API bill
| Model | Est. monthly cost |
|---|---|
| GPT-5.5 Global Standard | $110 |
| GPT-5.5 Data Zone Standard (base) | $121 |
| GPT-5.5 Global Standard, long context | $190 |
| GPT-5.5 Data Zone, long context (base) | $209 |
| GPT-5.4 Global Standard | $55 |
| GPT-5.4 Data Zone Standard (base) | $60.5 |
| GPT-5.4 Global Standard, long context | $95 |
| GPT-5.4 mini Global Standard | $16.5 |
Estimates use Azure OpenAI's standard per-token rates from the table above. For token-level presets and cross-provider comparison, use the full LLM API pricing calculator.
Azure OpenAI deployment types change the price
Global Standard is the default pay-per-token option: Azure can route a request across its global footprint and offers the highest quota. Data Zone Standard keeps processing inside a defined US, EU, or APAC data zone and carries a higher token meter in this snapshot. Regional Standard processes in one Azure region; its rate and model availability must be checked for that region.
Global Batchis the cost lane for asynchronous work. Microsoft documents it at 50% below Global Standard, with a 24-hour target completion window. The deployment behavior and current availability are described in Microsoft's deployment types guide.
Long context can move a request onto a higher meter
GPT-5.5 has separate short-context and long-context meters in Microsoft's retail feed. The Global Standard rate rises from $5/$30 to $10/$45 per million input/output tokens, while cached input rises from $0.50 to $1. The base Data Zone long-context meter is $11/$49.50. Do not budget a million-token workload from the headline short-context rate alone.
Microsoft labels the GPT-5.4 split explicitly: Global Standard is $2.50/$15 below 272K tokens and $5/$22.50 above 272K. If prompts regularly cross that line, trimming retrieved context can save more than changing deployment regions.
Azure OpenAI pricing vs the OpenAI API
Azure's GPT-5.5 Global Standard meter matches the direct OpenAI list rate in this snapshot: $5/$30 with $0.50 cached input. The base Data Zone meter is 10% higher at $5.50/$33. That premium buys Azure's data-zone routing and enterprise cloud integration; it does not represent a different underlying model.
Price is only one part of the choice. Azure adds deployment controls, regional availability, Azure identity, quota management, and provisioned capacity. Direct OpenAI can be operationally simpler. Compare the direct rates on the OpenAI API pricing page, then use the cross-provider cost calculator for the token mix.
Pay per token or reserve provisioned throughput?
Standard deployment is usually the safer starting point because cost scales with actual tokens. Provisioned deployments reserve processing capacity in Provisioned Throughput Units and bill per PTU-hour. The correct comparison is therefore not one token price against another; it is the measured throughput and utilization of a stable workload against the required PTU reservation.
Use Batch for delay-tolerant jobs, Global Standard for variable interactive traffic, Data Zone when the routing boundary is required, and evaluate provisioned throughput only after load tests show a sustained baseline. Microsoft documents the billing distinction in its provisioned throughput guide.
Azure OpenAI API pricing FAQ
How much does Azure OpenAI cost per million tokens?
Current Azure OpenAI prices depend on model and deployment. GPT-5.5 Global Standard is $5 input, $0.50 cached input, and $30 output per million tokens. The base Data Zone meter is $5.50/$33, and Global Batch is $2.50/$15.
Is Azure OpenAI more expensive than the OpenAI API?
Not always. GPT-5.5 Global Standard matches OpenAI's direct $5/$30 list rate in this snapshot. Azure Data Zone is higher at a base $5.50/$33, and other regions or contract terms can differ.
What is the difference between Global Standard and Data Zone pricing?
Global Standard can route inference globally. Data Zone Standard keeps processing inside a defined US, EU, or APAC zone. For GPT-5.5, the common Data Zone meter in Microsoft's retail feed is 10% above Global Standard, though some APAC meters are higher.
Does Azure OpenAI offer cheaper Batch pricing?
Yes. Microsoft documents Global Batch at 50% below Global Standard for asynchronous work with a 24-hour target completion window. GPT-5.5 Global Batch is $2.50/$15 per million input/output tokens.