SunsetDiff

Azure OpenAI gpt-5.3-codex, gpt-audio-1.5, and gpt-realtime-1.5 retirement on August 24, 2027

The dataset has 3 Azure OpenAI ids with shutdown date 2027-08-24: gpt-5.3-codex-2026-02-24, gpt-audio-1.5-2026-02-23, gpt-realtime-1.5-2026-02-23. The retirement schedule lists that date on each of those rows. Lifecycle is GA. The replacement cell is empty, so those cells say Not published. The lifecycle page says a Retired model is removed from service and all inference requests return 410 Gone. gpt-5.3-codex version 2026-02-24 is a Responses API model used by Codex. gpt-audio-1.5 version 2026-02-23 is an audio model for audio and text generation. gpt-realtime-1.5 version 2026-02-23 is an audio model for real-time audio processing. OpenAI's API retires gpt-5.3-codex on 2027-04-01. Microsoft's Azure schedule does not print that date for these rows.

Last reviewed . Publisher: SunsetDiff.

Steps

  1. Find deployments. Call Deployments List, or Get-AzCognitiveServicesAccountDeployment, for each Azure OpenAI resource. Read properties.model.name and properties.model.version. A deployment name can differ from the model name. gpt-5.3-codex version 2026-02-24 is the dataset id gpt-5.3-codex-2026-02-24. gpt-audio-1.5 version 2026-02-23 is gpt-audio-1.5-2026-02-23. gpt-realtime-1.5 version 2026-02-23 is gpt-realtime-1.5-2026-02-23.
  2. Pick a replacement. The replacement cell is empty for gpt-5.3-codex 2026-02-24, gpt-audio-1.5 2026-02-23, and gpt-realtime-1.5 2026-02-23. Microsoft did not name a recommended replacement on these rows. Shortlist a model yourself and test it. Do not copy OpenAI's gpt-6-sol onto the Azure gpt-5.3-codex row. The Codex how-to sample sets the model string to gpt-5-codex and says to replace it with your deployment name. That sample string is not the replacement cell.
  3. Test. Send one real request to the model you picked before August 24, 2027. For gpt-5.3-codex, send a coding prompt on the Responses API. For gpt-realtime-1.5, send one realtime audio call and check latency, the voice, and pcm16. For gpt-audio-1.5, send one audio and text request. The audio quickstart does not name gpt-audio-1.5 in its supported-models table, so confirm the API version and the voice on the call you already send.
  4. Redeploy before the date. Create a deployment of the model you tested and move traffic before August 24, 2027. These rows do not name an automatic-upgrade target. Provisioned deployments are not auto-upgraded. A deployment set to NoAutoUpgrade stops working on the retirement date.

Models and dates

These 3 ids are the Azure OpenAI rows whose shutdown date is 2027-08-24. The model id links to its SunsetDiff model page. Lifecycle is the schedule's Lifecycle column, copied from the source excerpt stored on the row. Shutdown date and replacement are the fields stored on the row. Not published means the replacement field is empty. These rows do not store a deprecation_date. gpt-5.3-codex-2026-02-24 (GA), gpt-audio-1.5-2026-02-23 (GA), gpt-realtime-1.5-2026-02-23 (GA).

Checked on 2026-11-05, the retirement schedule lists gpt-5.3-codex version 2026-02-24, gpt-audio-1.5 version 2026-02-23, and gpt-realtime-1.5 version 2026-02-23 as GA with retirement date 2027-08-24 and an empty replacement cell. Those rows do not carry footnote 2 or footnote 3. The schedule does not print a second row for any of these three versions.

The lifecycle table says Generally Available means: "Production-ready. Weights and APIs are fixed." New deployments are allowed, and existing deployments work, until the retirement date. The lifecycle page also says: "Check the Model retirement schedule for the current retirement dates and replacement model, since these details are subject to change."

The Azure Government schedule does not list gpt-5.3-codex, gpt-audio-1.5, or gpt-realtime-1.5. The dates in the first table are the public-cloud schedule.

3 Azure OpenAI model ids with shutdown date 2027-08-24.
Model idLifecycle on the scheduleShutdown dateRecommended replacement
gpt-5.3-codex-2026-02-24GA2027-08-24Not published
gpt-audio-1.5-2026-02-23GA2027-08-24Not published
gpt-realtime-1.5-2026-02-23GA2027-08-24Not published

Sources: Azure OpenAI retirement schedule , Foundry Models lifecycle , Azure Government retirement schedule , checked November 5, 2026.

Which call each model uses

The models page lists gpt-5.3-codex (2026-02-24). The description cell includes Reasoning, Responses API, Structured outputs, Text and image processing, Functions, tools, and parallel tool calling, Full summary of capabilities, and Optimized for Codex CLI & Codex VS Code extension. The context window cell lists 400,000, with input 272,000 and output 128,000. The max output tokens cell is 128,000. The training data cell says August 2025.

The Codex how-to says: "From the model catalog select a reasoning model such as gpt-6-astra, gpt-6-sol, gpt-5.3-codex, gpt-5.2-codex, gpt-5.1-codex-max, gpt-5.1-codex, gpt-5.1-codex-mini, gpt-5-codex, gpt-5, gpt-5-mini, or gpt-5-nano." That sentence is a list of reasoning models you can deploy. It is not the Replacement column on the retirement schedule. The same page says: "With the v1 API you no longer need to pass api-version, but you must include /v1 in the base_url path." The sample sets wire_api to responses and the model string to gpt-5-codex. The comment on that line says: "Replace with your actual Azure model deployment name" Put the deployment name in that field. The sample string gpt-5-codex is not the replacement for gpt-5.3-codex version 2026-02-24.

The models page says: "Audio models in Azure OpenAI are available via the realtime, completions, and audio APIs." Under GPT-4o audio models it says: "The GPT-4o audio models are part of the GPT-4o model family and support either low-latency, speech in, speech out conversational interactions or audio generation." The table under that sentence lists gpt-audio-1.5 (2026-02-23) as "Audio model for audio and text generation." The max request tokens cell prints input 128,000 and output 16,384. Training data is up to September 2024. It lists gpt-realtime-1.5 (2026-02-23) as "Audio model for real-time audio processing." The max request tokens cell prints input 32,000 and output 4,096. Training data is up to September 2024.

What's new says: "The gpt-realtime-1.5 and gpt-audio-1.5 models are now available." Then: "These models build on last year's GPT-Realtime and GPT-Audio with improvements in instruction following, multilingual support, and tool calling, while preserving the low-latency, real-time interactions developers need for voice-based applications." Then: "Try them with the existing chat completion APIs in Microsoft Foundry." That page does not print an api-version for either model.

The realtime how-to lists gpt-realtime-1.5 (2026-02-23) among its supported models. It says the API supports low-latency, speech in, speech out conversational interactions. It says: "For all Realtime API models, use the GA endpoint format with /openai/v1 in the URL." And: "For the Realtime API, use the GA endpoint with /openai/v1 in the URL. Don't use date-based API versions or the api-version query parameter." It also says: "The Realtime API supports up to 32,000 input tokens and 4,096 output tokens." The audio quickstart supported-models table includes gpt-realtime-1.5. The primary use cell is real-time interactive voice. That table does not include gpt-audio-1.5.

gpt-5.3-codex, gpt-audio-1.5, and gpt-realtime-1.5 are three calls. The schedule does not name a model that keeps the same call. When you shortlist a target, confirm that target still accepts the call you send now.

Sources: Foundry Models sold by Azure , Codex with Azure OpenAI , Azure OpenAI what's new , GPT Realtime API , Azure OpenAI audio generation quickstart , checked November 5, 2026.

Replacements Microsoft names

The schedule says: "The Replacement column identifies Microsoft's recommended replacement and, where supported, the target for automatic upgrades at retirement." It also says: "You don't need to wait for a replacement to be listed before evaluating or manually migrating to another compatible model." The lifecycle page says: "The replacement listed in the Model retirement schedule is Microsoft's recommended migration target and, where supported, the model used for automatic upgrades at retirement. It doesn't limit your manual migration choices."

No row in this table stores a replacement.

The replacement cell is empty on gpt-5.3-codex-2026-02-24, gpt-audio-1.5-2026-02-23, gpt-realtime-1.5-2026-02-23. Empty means Microsoft did not name a recommended replacement on that row. This guide does not add one.

The lifecycle page says: "When you choose a target, evaluate candidate models against your own application, prompts, and representative data. Compare quality, latency, and cost together rather than relying on public benchmarks alone." Then: "Before you migrate, confirm API compatibility, feature support, deployment-type and regional availability, capacity, and quota."

The Codex deploy list names gpt-6-sol. The retirement schedule does not print gpt-6-sol in the replacement cell for gpt-5.3-codex version 2026-02-24. This guide does not turn that deploy list into an Azure replacement.

Sources: Azure OpenAI retirement schedule , Foundry Models lifecycle , Codex with Azure OpenAI , checked November 5, 2026.

What to test

Send one real request before August 24, 2027. The lifecycle page says: "When you choose a target, evaluate candidate models against your own application, prompts, and representative data. Compare quality, latency, and cost together rather than relying on public benchmarks alone." That sentence names latency. These pages do not print one latency number that covers all 3 models.

gpt-5.3-codex version 2026-02-24: the models page lists the Responses API, structured outputs, text and image processing, and tools. The Codex sample uses /openai/v1 and does not pass api-version. Test one coding prompt you send now, on the deployment name you will keep. The context window cell lists input 272,000 and output 128,000.

gpt-realtime-1.5 version 2026-02-23: the realtime how-to says: "In most cases, we recommend using the WebRTC API for low-latency real-time audio streaming." Its connection-methods table lists WebRTC at ~100ms, WebSocket at ~200ms, and SIP as Varies. The example session uses voice alloy and input_audio_format pcm16. The page says: "The Realtime API expects audio in a specific format:" The list under that sentence is PCM 16-bit (pcm16), mono (single channel), and 24kHz. It also says: "Realtime sessions have a maximum duration of 60 minutes." The models page lists input 32,000 and output 4,096 for this version. The how-to says: "For the Realtime API, use the GA endpoint with /openai/v1 in the URL. Don't use date-based API versions or the api-version query parameter."

gpt-audio-1.5 version 2026-02-23: the models page calls it an audio model for audio and text generation, with input 128,000 and output 16,384. What's new says to try it with the existing chat completion APIs. The audio quickstart says: "The following voices are supported for audio out: Alloy, Ash, Ballad, Coral, Echo, Sage, Shimmer, Verse, Marin, and Cedar." And: "The following audio output formats are supported: wav, mp3, flac, opus, pcm16, and aac." And: "Support for audio completions was first added in API version 2025-01-01-preview." The supported-models table on that page does not include gpt-audio-1.5, and the sample sets the model to gpt-audio-mini. This guide does not copy that voice list, those formats, or API version 2025-01-01-preview onto gpt-audio-1.5. Confirm the API version and the voice on the call you already send.

Sources: Foundry Models sold by Azure , Codex with Azure OpenAI , Azure OpenAI what's new , GPT Realtime API , Azure OpenAI audio generation quickstart , Foundry Models lifecycle , checked November 5, 2026.

OpenAI's date is earlier for gpt-5.3-codex

OpenAI's API deprecations page says: "The following models are deprecated and will be removed from the API on April 1, 2027, with six months' notice. Migrate to the recommended replacements before the shutdown date." That date is 2027-04-01. The same table lists gpt-5.3-codex with replacement gpt-6-sol. The OpenAI row is on GPT-5.1, gpt-5.3-codex, and gpt-5.4-nano shutdown on April 1, 2027. gpt-5.3-codex stores shutdown date 2027-04-01 and replacement gpt-6-sol.

Microsoft's retirement schedule and lifecycle page do not print 2027-04-01 as a shutdown date for Azure gpt-5.3-codex version 2026-02-24. The Azure row stays 2027-08-24, and its replacement cell is empty. A team that calls the OpenAI API and Azure should plan the OpenAI gpt-5.3-codex calls for 2027-04-01. The Azure deployment keeps 2027-08-24. Do not copy gpt-6-sol onto the Azure row.

The same OpenAI page names gpt-audio-1.5 as a replacement for older audio ids, including gpt-audio and gpt-audio-mini. It names gpt-realtime-1.5 as a replacement for gpt-4o-realtime-preview. It does not print a shutdown date for gpt-audio-1.5 or gpt-realtime-1.5. gpt-audio-1.5 stores shutdown date Not published and replacement Not published. gpt-realtime-1.5 stores shutdown date Not published and replacement Not published. Those stored rows are the OpenAI API. The Azure rows still use 2027-08-24.

Sources: Azure OpenAI retirement schedule , Foundry Models lifecycle , OpenAI deprecations , checked November 5, 2026.

Other Azure rows are a different date

June 15, 2027 is an earlier audio date. gpt-4o-mini-transcribe-2025-12-15 stores shutdown date 2027-06-15 and replacement Not published. gpt-4o-mini-tts-2025-12-15 stores shutdown date 2027-06-15 and replacement Not published. gpt-audio-mini-2025-12-15 stores shutdown date 2027-06-15 and replacement Not published. gpt-realtime-mini-2025-12-15 stores shutdown date 2027-06-15 and replacement Not published. The guide is Azure OpenAI gpt-4o-mini-tts, transcribe, and realtime audio retirement on June 15, 2027. The replacement cell on those rows is empty. They do not name gpt-audio-1.5 or gpt-realtime-1.5.

These other stored rows are not the 2027-08-24 table. gpt-audio-2025-08-28 stores shutdown date 2027-03-02 and replacement Not published. gpt-realtime-2025-08-28 stores shutdown date 2027-03-02 and replacement Not published. gpt-5-codex-2025-09-15 stores shutdown date 2027-03-17 and replacement Not published. gpt-realtime-2-2026-05-06 stores shutdown date 2027-01-20 and replacement gpt-realtime-2.1. gpt-realtime-2.1-2026-07-07 stores shutdown date 2027-06-25 and replacement Not published. gpt-realtime-2.1-mini-2026-07-07 stores shutdown date 2027-06-25 and replacement Not published. gpt-realtime-2.1 stores shutdown date Not published and replacement Not published.

The schedule names gpt-realtime-2.1 as the replacement for gpt-realtime-2 version 2026-05-06. It does not name gpt-realtime-2.1 as the replacement for gpt-realtime-1.5 version 2026-02-23. The dated row gpt-realtime-2.1-2026-07-07 stores shutdown date 2027-06-25, which is before 2027-08-24. The Codex sample string gpt-5-codex is the family of gpt-5-codex-2025-09-15, and that stored row retires on 2027-03-17. Neither date is an Azure replacement for the rows in the first table.

Sources: Azure OpenAI retirement schedule , Foundry Models lifecycle , checked November 5, 2026.

What happens to a deployment on the date

The lifecycle table says Retired means: "Removed from service. All inference requests return 410 Gone." New deployments are not allowed, and existing deployments do not work. The same page says the API value lifecycleStatus Deprecated means the model is retired and inference returns 410 Gone. The docs stage Deprecated is lifecycleStatus Deprecating, and inference still works. The sentences and the status map are on Azure OpenAI retired models return 410 Gone. These three rows are GA now, so inference still works until August 24, 2027.

The lifecycle page says: "For Global Standard, Data Zone Standard, and Standard deployment types, Microsoft manages automatic upgrades when a model version is retired:" It also says: "Provisioned deployments are NOT auto-upgraded. Provisioned customers must manually migrate to the replacement model." The migration page says: "Standard, Global Standard, and Data Zone Standard (pay-as-you-go) deployments are auto-upgraded on a rolling, region-by-region schedule." And: "Provisioned (PTU) deployments are not auto-upgraded." And: "Batch deployments follow the side-by-side path: deploy the new model, resubmit jobs, retire the old deployment."

Working with models says: "Automatic model updates are only supported for Standard deployment types." OnceCurrentVersionExpired means: "Once the retirement date is reached the model deployment automatically upgrades to the current default version." NoAutoUpgrade means: "The model deployment never automatically upgrades. Once the retirement date is reached the model deployment stops working." And: "null is equivalent to OnceCurrentVersionExpired."

The same page says: "When you select a specific model version for a deployment, this version remains selected until you either choose to manually update yourself, or once you reach the retirement date for the model." Then: "When the retirement date is reached the model will automatically upgrade to the default version at the time of retirement." A deployment you pinned to one of these versions can still move on the retirement date unless versionUpgradeOption is NoAutoUpgrade. NoAutoUpgrade does not move, and the deployment stops. The replacement cell is the target where an automatic upgrade is supported. These three rows have an empty cell, so they do not name that target.

Sources: Azure OpenAI retirement schedule , Foundry Models lifecycle , Working with Azure OpenAI models , Model migration process , checked November 5, 2026.

Find deployments that use these versions

The migration page says: "You also need an inventory of the model deployments your workload uses, including their deployment types (Standard, Provisioned, or Batch)." Working with models says: "To query the current model deployment settings including the deployment upgrade configuration for a given resource use Deployments List." If the value is null, you won't see a versionUpgradeOption property. The model versions page says: "Go to your deployment in the Foundry portal, then open the Details tab to see the current model version." And: "Query the deployments endpoint for your resource. The response includes the model version for each deployment."

The example response puts the model name and version on properties.model. The deployment's own name is the name field next to that object. Match properties.model.name and properties.model.version to the ids in the table. gpt-5.3-codex version 2026-02-24 is the dataset id gpt-5.3-codex-2026-02-24.

Also search the repository. A hit is a candidate, not proof a deployment still uses that version. How to find every model ID in your codebase covers SDK argument names. The scan page builds a search from dataset ids that have a shutdown date.

Azure PowerShell, list deployments for one account

Get-AzCognitiveServicesAccountDeployment -ResourceGroupName {ResourceGroupName} -AccountName {AccountName}

Deployments List

GET https://management.azure.com/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.CognitiveServices/accounts/{accountName}/deployments?api-version=2023-05-01

ripgrep

rg -n -w -e 'gpt-5.3-codex-2026-02-24' -e 'gpt-audio-1.5-2026-02-23' -e 'gpt-realtime-1.5-2026-02-23'

Sources: Working with Azure OpenAI models , Model versions , Model migration process , checked November 5, 2026.

Fine-tuned training and deployment

The schedule and the lifecycle page both say: "Fine-tuned models retire in two phases: training and deployment." And: "Unless explicitly stated, training retires no earlier than the base model retirement date." Both pages say: "For existing customers only. Otherwise, training retirement occurs at base model retirement."

The fine-tune tables on both pages do not list gpt-5.3-codex version 2026-02-24, gpt-audio-1.5 version 2026-02-23, or gpt-realtime-1.5 version 2026-02-23. gpt-5.3-codex-2026-02-24: no fine-tune dates are stored, and the fine-tune tables do not list this version. gpt-audio-1.5-2026-02-23: no fine-tune dates are stored, and the fine-tune tables do not list this version. gpt-realtime-1.5-2026-02-23: no fine-tune dates are stored, and the fine-tune tables do not list this version.

A base deployment of these versions uses the retirement date in the first table. Do not borrow a fine-tune date from a different model.

Sources: Azure OpenAI retirement schedule , Foundry Models lifecycle , checked November 5, 2026.

Redeploy before August 24, 2027

Create a deployment of a model you have tested, point traffic at it, and confirm one real request before August 24, 2027. Do that for each of the 3 versions in the table.

Working with models describes two ways to change a provisioned deployment once you have chosen a target. In-place migration keeps the deployment name and moves traffic. Multi-deployment migration creates a second deployment, shifts traffic, then deletes the old one. Neither section names a target for these rows.

The batch of these rows is Azure OpenAI shutdowns on August 24, 2027. The Azure OpenAI provider page lists the other stored rows. The earlier audio batch is Azure OpenAI gpt-4o-mini-tts, transcribe, and realtime audio retirement on June 15, 2027. The July 13, 2027 batch is Azure OpenAI gpt-5.1, gpt-5-mini, and gpt-5.2-codex retirement on July 13, 2027.

The pre-migration test checklist is the longer list: evals, rejected fields, token limits, tools, price, and latency. These Azure rows do not store a price.

Sources: Model migration process , Working with Azure OpenAI models , Azure OpenAI retirement schedule , checked November 5, 2026.

Related SunsetDiff pages

All guides

Questions

Which Azure OpenAI versions retire on August 24, 2027?

The dataset lists gpt-5.3-codex-2026-02-24, gpt-audio-1.5-2026-02-23, gpt-realtime-1.5-2026-02-23. The retirement schedule lists shutdown date 2027-08-24 for each of them. gpt-5.3-codex-2026-02-24 (GA), gpt-audio-1.5-2026-02-23 (GA), gpt-realtime-1.5-2026-02-23 (GA). gpt-5.3-codex version 2026-02-24, gpt-audio-1.5 version 2026-02-23, and gpt-realtime-1.5 version 2026-02-23.

What is the official replacement?

The recommended replacement is the official replacement stored on each row. gpt-5.3-codex-2026-02-24: Not published. gpt-audio-1.5-2026-02-23: Not published. gpt-realtime-1.5-2026-02-23: Not published. The schedule says: "The Replacement column identifies Microsoft's recommended replacement and, where supported, the target for automatic upgrades at retirement." The cell is empty on gpt-5.3-codex-2026-02-24, gpt-audio-1.5-2026-02-23, gpt-realtime-1.5-2026-02-23. This guide does not add gpt-6-sol, gpt-5-codex, or gpt-realtime-2.1 as an Azure replacement for these rows.

Does OpenAI's April 1, 2027 shutdown of gpt-5.3-codex apply to Azure?

No for the Azure deployment. Microsoft's retirement schedule and lifecycle page do not print 2027-04-01 as an Azure gpt-5.3-codex shutdown. The schedule lists Azure gpt-5.3-codex version 2026-02-24 as GA with retirement date 2027-08-24 and an empty replacement cell. OpenAI's deprecations page says: "The following models are deprecated and will be removed from the API on April 1, 2027, with six months' notice. Migrate to the recommended replacements before the shutdown date." The same table lists gpt-5.3-codex with replacement gpt-6-sol. That sentence is the OpenAI API. See GPT-5.1, gpt-5.3-codex, and gpt-5.4-nano shutdown on April 1, 2027. gpt-5.3-codex stores shutdown date 2027-04-01 and replacement gpt-6-sol.

What call does each model use?

gpt-5.3-codex-2026-02-24 is on the Responses API. The models page lists that API, and the Codex sample uses /openai/v1 with wire_api responses. gpt-audio-1.5-2026-02-23 is an audio model for audio and text generation. What's new says: "Try them with the existing chat completion APIs in Microsoft Foundry." The audio quickstart supported-models table does not include gpt-audio-1.5. gpt-realtime-1.5-2026-02-23 is real-time audio. The realtime how-to lists version 2026-02-23 and says: "For the Realtime API, use the GA endpoint with /openai/v1 in the URL. Don't use date-based API versions or the api-version query parameter."

What should I test before August 24, 2027?

Test one real call before 2027-08-24. For gpt-5.3-codex version 2026-02-24, send a coding prompt on the Responses API and check tools and structured outputs. For gpt-realtime-1.5 version 2026-02-23, test latency, the voice, and pcm16. The realtime connection table lists WebRTC at ~100ms, WebSocket at ~200ms, and SIP as Varies. Input is PCM 16-bit (pcm16), mono, 24kHz. Sessions last at most 60 minutes. For gpt-audio-1.5 version 2026-02-23, send one audio and text request. The quickstart does not print an api-version for that id.

Do gpt-audio-1.5 and gpt-realtime-1.5 have an OpenAI shutdown date?

OpenAI's deprecations page does not print a shutdown date for gpt-audio-1.5 or gpt-realtime-1.5. It names them as replacements for older audio and realtime ids. gpt-audio-1.5 stores shutdown date Not published and replacement Not published. gpt-realtime-1.5 stores shutdown date Not published and replacement Not published. The Azure rows gpt-audio-1.5-2026-02-23 and gpt-realtime-1.5-2026-02-23 store 2027-08-24.

What does Azure return after retirement?

The lifecycle page says a Retired model is removed from service and: "Removed from service. All inference requests return 410 Gone." See Azure OpenAI retired models return 410 Gone.

Will every deployment upgrade itself?

No. The lifecycle page says: "For Global Standard, Data Zone Standard, and Standard deployment types, Microsoft manages automatic upgrades when a model version is retired:" And: "Provisioned deployments are NOT auto-upgraded. Provisioned customers must manually migrate to the replacement model." Working with models says: "Automatic model updates are only supported for Standard deployment types." NoAutoUpgrade means: "The model deployment never automatically upgrades. Once the retirement date is reached the model deployment stops working." null is equivalent to OnceCurrentVersionExpired. These three rows have an empty replacement cell, so they do not name an upgrade target.

How do I find deployments that still use these versions?

Working with models says: "To query the current model deployment settings including the deployment upgrade configuration for a given resource use Deployments List." Read properties.model.name and properties.model.version. The model versions page says: "Go to your deployment in the Foundry portal, then open the Details tab to see the current model version." The migration page says: "You also need an inventory of the model deployments your workload uses, including their deployment types (Standard, Provisioned, or Batch)." gpt-5.3-codex version 2026-02-24 is gpt-5.3-codex-2026-02-24. gpt-audio-1.5 version 2026-02-23 is gpt-audio-1.5-2026-02-23. gpt-realtime-1.5 version 2026-02-23 is gpt-realtime-1.5-2026-02-23.

Do fine-tuned deployments of these versions retire on August 24, 2027?

The base retirement date is 2027-08-24. The fine-tune tables do not list these three versions. gpt-5.3-codex-2026-02-24: no fine-tune dates are stored, and the fine-tune tables do not list this version. gpt-audio-1.5-2026-02-23: no fine-tune dates are stored, and the fine-tune tables do not list this version. gpt-realtime-1.5-2026-02-23: no fine-tune dates are stored, and the fine-tune tables do not list this version. Both pages say: "For existing customers only. Otherwise, training retirement occurs at base model retirement."

Does Azure Government list the same date?

The Azure Government schedule does not list gpt-5.3-codex, gpt-audio-1.5, or gpt-realtime-1.5. The public-cloud schedule lists all three ids on 2027-08-24.

What about the June 15, 2027 audio models, gpt-realtime-2.1, and the Codex sample string?

gpt-4o-mini-transcribe-2025-12-15 stores shutdown date 2027-06-15 and replacement Not published. gpt-4o-mini-tts-2025-12-15 stores shutdown date 2027-06-15 and replacement Not published. gpt-audio-mini-2025-12-15 stores shutdown date 2027-06-15 and replacement Not published. gpt-realtime-mini-2025-12-15 stores shutdown date 2027-06-15 and replacement Not published. See Azure OpenAI gpt-4o-mini-tts, transcribe, and realtime audio retirement on June 15, 2027. gpt-audio-2025-08-28 stores shutdown date 2027-03-02 and replacement Not published. gpt-realtime-2025-08-28 stores shutdown date 2027-03-02 and replacement Not published. gpt-5-codex-2025-09-15 stores shutdown date 2027-03-17 and replacement Not published. gpt-realtime-2-2026-05-06 stores shutdown date 2027-01-20 and replacement gpt-realtime-2.1. gpt-realtime-2.1-2026-07-07 stores shutdown date 2027-06-25 and replacement Not published. gpt-realtime-2.1-mini-2026-07-07 stores shutdown date 2027-06-25 and replacement Not published. gpt-realtime-2.1 stores shutdown date Not published and replacement Not published. gpt-realtime-2 version 2026-05-06 retires on 2027-01-20 and names gpt-realtime-2.1. See Azure OpenAI shutdowns on January 20, 2027 and Azure OpenAI retirements from January 12 to February 9, 2027. gpt-realtime-2.1 version 2026-07-07 retires on 2027-06-25. gpt-5-codex version 2025-09-15 retires on 2027-03-17. None of those cells is the replacement for the 2027-08-24 rows.

More guides

Deprecation alerts

We email you before models shut down. Unsubscribe anytime.

Deprecation alert signup
Providers (optional)