{"slug": "azure-openai-model-retirement-every-date-auto-upgrades-and-how-to-migrate", "title": "Azure OpenAI model retirement: every date, auto-upgrades and how to migrate", "summary": "Microsoft's Azure OpenAI retirement schedule sets gpt-4o version 2024-05-13 to retire on January 14, 2027, when Global Standard and Data Zone Standard deployments are automatically upgraded in place to gpt-5.1 version 2025-11-13 and Provisioned deployments stop serving. On April 14, 2027, Azure retires gpt-4o versions 2024-08-06 and 2024-11-20, gpt-4o-mini, gpt-4.1, gpt-4.1-mini and gpt-4.1-nano, while the o-series models o1, o1-pro, o3, o3-pro, o3-mini, o4-mini and o3-deep-research were scheduled to retire on November 19, 2026. Microsoft's FAQ states retirement dates are not extendable, and the official replacement is named about 90 to 120 days before retirement.", "body_md": "[← all posts](https://www.evalshift.dev/blog)\n\n# Azure OpenAI model retirement: every date, auto-upgrades and how to migrate\n\nAzure OpenAI retires gpt-4o 2024-05-13 on January 14, 2027 and GPT-4o mini and GPT-4.1 on April 14. Every date, what auto-upgrades, and how to migrate.\n\nThe next big Azure OpenAI model retirement is `gpt-4o` version 2024-05-13 on January 14, 2027. On\nthat day, Global Standard and Data Zone Standard deployments of it are upgraded to `gpt-5.1`\nversion 2025-11-13 automatically, and Provisioned deployments stop serving. Three months later, on\nApril 14, 2027, Azure retires `gpt-4o` versions 2024-08-06 and 2024-11-20, `gpt-4o-mini`,\n`gpt-4.1`, `gpt-4.1-mini` and `gpt-4.1-nano`. Every o-series model still on Azure's schedule (` o1`,\n`o1-pro`, `o3`, `o3-pro`, `o3-mini`, `o4-mini` and `o3-deep-research`) was scheduled to retire on\nNovember 19, 2026.\n\nFour things about Azure's retirements catch teams that know OpenAI's:\n\n- +**The dates are Azure's own.**`gpt-4o` 2024-08-06 shuts down on OpenAI's API on January 14,\n2027 and on Azure on April 14, 2027. The o-series is mixed:`o3` and`o3-pro` were scheduled to\nretire on Azure on November 19, 2026, three weeks before OpenAI's December 11, while`o1` ,`o3-mini` and`o4-mini` shut down on OpenAI's API first, on October 23.\n- +**A retirement can swap the model under you.** Standard deployments are upgraded at retirement\nby default, under the same deployment name. A Global or Data Zone Standard deployment of`gpt-4o` 2024-05-13 is upgraded in place, region by region, from January 14, and starts answering with a\nreasoning model your code never called.\n- +**Deprecated means closed to new subscriptions.** Once a model is Deprecated, only an Azure\nsubscription that has already deployed that exact version can create new deployments of it. A\nnew subscription in the same tenant can't.\n- +**The replacement is named late, and the date can't move.** Microsoft picks the official\nreplacement about 90 to 120 days before retirement, and its FAQ says retirement dates aren't\nextendable.\n\nThe rest of this post is every upcoming Azure OpenAI date for text and reasoning models, what\nhappens to each deployment type on the day, what changes when `gpt-4o` becomes `gpt-5.1`, the\nfine-tuned model dates, and a six-step migration. The facts come from Microsoft's\n[model retirement schedule ↗](https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule)\n(last updated November 18, 2026), its\n[lifecycle and support policy ↗](https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirements),\nits [model versions ↗](https://learn.microsoft.com/en-us/azure/foundry/foundry-models/concepts/model-versions)\nand [reasoning models ↗](https://learn.microsoft.com/en-us/azure/foundry/openai/how-to/reasoning)\npages, all read on November 19, 2026. Azure OpenAI is now documented under Microsoft Foundry, and\nMicrosoft revises the schedule often, so check it again before you act on a date.\n\n## Azure OpenAI retirement dates, November 2026 to July 2027\n\nThese are the chat, reasoning and Codex models. Each row is one model version, which is how Azure retires models: a deployment pins a version, and the version is what retires.\n\n| Retirement date | Model and version | Stage today | Listed replacement | \n|---|---|---|---|\n| November 17, 2026 (past) | `codex-mini` 2025-05-16;`gpt-chat-latest` 2026-05-05 and 2026-06-24 | Retired | None listed | \n| November 19, 2026 (past) | `o1` 2024-12-17,`o1-pro` 2025-03-19,`o3` 2025-04-16,`o3-pro` 2025-06-10,`o3-deep-research` 2025-06-26 | Retired | `gpt-5.6-sol` | \n| November 19, 2026 (past) | `o3-mini` 2025-01-31,`o4-mini` 2025-04-16 | Retired | `gpt-5.6-terra` | \n| December 8, 2026 | `gpt-chat-latest` 2026-05-28 | Preview | None listed | \n| January 12, 2027 | `gpt-chat-latest` 2026-08-06 | Preview | None listed | \n| January 14, 2027 | `gpt-4o` 2024-05-13 | Deprecated | `gpt-5.1` 2025-11-13, applied automatically on Global Standard and Data Zone Standard | \n| February 2, 2027 | `gpt-chat-latest` 2026-10-07 | Preview | None listed | \n| February 9, 2027 | `gpt-5` 2025-08-07,`gpt-5-nano` 2025-08-07 | GA | None listed | \n| March 17, 2027 | `gpt-5-codex` 2025-09-15 | GA | None listed | \n| April 7, 2027 | `gpt-5-pro` 2025-10-06 | GA | None listed | \n| April 14, 2027 | `gpt-4o` 2024-08-06 and 2024-11-20 | Deprecated; Legacy | `gpt-5.1` | \n| April 14, 2027 | `gpt-4o-mini` 2024-07-18,`gpt-4.1` ,`gpt-4.1-mini` and`gpt-4.1-nano` 2025-04-14 | Deprecated | None listed yet | \n| May 15, 2027 | `gpt-5.1-codex` and`gpt-5.1-codex-mini` 2025-11-13 | GA | None listed | \n| May 18, 2027 | `gpt-5.1-codex-max` 2025-12-04 | GA | None listed | \n| June 8, 2027 | `gpt-5.2` 2025-12-11 | GA | None listed | \n| July 13, 2027 | `gpt-5-mini` 2025-08-07,`gpt-5.1` 2025-11-13,`gpt-5.2-codex` 2026-01-14 | GA | None listed | \n\n\"None listed yet\" on the April 14 row is expected. Microsoft names the official replacement about\n90 to 120 days before a retirement, so the `gpt-4o-mini` and GPT-4.1 targets should appear\nbetween mid-December and mid-January. Don't wait for it: the policy says the listed replacement\n\"doesn't limit your manual migration choices.\"\n\nThe audio and speech models have their own dates. `whisper`, `tts` and `tts-hd` retire on\nDecember 15, 2026, the March 2025 `gpt-4o-mini-transcribe` and `gpt-4o-mini-tts` versions on\nJanuary 20, 2027, `gpt-4o-transcribe` 2025-03-20 on February 26, 2027, and `gpt-audio` and\n`gpt-realtime` 2025-08-28 on March 2, 2027. The embedding models, `text-embedding-3-large`,\n`text-embedding-3-small` and `text-embedding-ada-002`, run until February 9, 2028.\n\nFor the same dates side by side with OpenAI, Anthropic, Google and AWS Bedrock, see the\n[LLM model deprecation schedule](https://www.evalshift.dev/blog/llm-model-deprecation-schedule).\n\n## What happens to a deployment on the retirement date\n\nA retired model version returns `410 Gone` on every inference request. Whether your deployment is\nstill on that version when the date arrives depends on two things: its deployment type and its\nversion upgrade policy.\n\n| Deployment | What happens at retirement | \n|---|---|\n| Global Standard, Data Zone Standard or Standard, upgrade policy `OnceCurrentVersionExpired` (the default) | Upgraded under the same deployment name, on a rolling region-by-region schedule: to Microsoft's listed replacement where it's offered for that deployment type and region, otherwise to the current default version | \n| Same, upgrade policy `OnceNewDefaultVersionAvailable` | Already moves to each new default version within two weeks of it becoming the default | \n| Same, upgrade policy `NoAutoUpgrade` | Stops working. Requests fail until you point the deployment at a supported version | \n| Global, Data Zone or Regional Provisioned | Never upgraded automatically. Stops serving unless you migrated it first | \n| Global Batch | Not covered by the automatic upgrade. Deploy the new model and resubmit jobs to it | \n| Preview models, any type | Force-upgraded to a newer version or retired, with at least 30 days' notice. There's no staying on a retiring preview | \n\nTwo details in the upgrade path matter. First, a deployment that never had its policy set shows\nno `versionUpgradeOption` at all, and Microsoft documents a missing value as equivalent to\n`OnceCurrentVersionExpired`. So unless someone chose otherwise, your Standard deployments will be\nupgraded. Second, the upgrade doesn't wait for the replacement to exist in your region. Microsoft's\nmodel versions page says that if the new version isn't available in a region yet, it deploys it\nthere as part of the upgrade.\n\n`gpt-4o` 2024-05-13 has a third wrinkle: its Regional Standard deployments already retired on March\n31, 2026. The January 14 date covers Global Standard, Data Zone Standard, the three Provisioned\ntypes and Global Batch.\n\n## What changes when gpt-4o becomes gpt-5.1 on January 14\n\nMicrosoft's listed replacement for every `gpt-4o` version is `gpt-5.1`, pinned to version\n2025-11-13 for the 2024-05-13 retirement. For that version, Global Standard and Data Zone Standard deployments switch to it automatically\non January 14. The deployment keeps its name, your code keeps calling the same endpoint, and the\nanswers now come from a different model. Here is what differs:\n\n- +**`gpt-5.1` is a reasoning model.** On Azure its`reasoning_effort` defaults to`none` , so after\nthe upgrade it answers without reasoning unless you send a higher effort. Microsoft's reasoning\npage calls this out as a migration note for code coming from older reasoning models.\n- +**Some request parameters are listed as unsupported.** Azure's reasoning models page says\nreasoning models other than GPT-6 Astra don't support`temperature` ,`top_p` ,`presence_penalty` ,`frequency_penalty` ,`logprobs` ,`top_logprobs` ,`logit_bias` or`max_tokens` , and that Chat Completions needs`max_completion_tokens` instead of`max_tokens` .\nThe page makes no exception for`none` , so test the exact request your code sends: a`gpt-4o` request with`temperature` and`max_tokens` is the most common shape there is.\n- +**Azure's feature table lists streaming as unsupported for `gpt-5.1` 2025-11-13.** OpenAI's own\nmodel page lists streaming as supported, so this may be a gap in Azure's table, but test your\nstreaming requests against a`gpt-5.1` deployment before January 14 rather than find out on the\nday.\n- +**Reasoning tokens are billed as output.** At`none` there are few or none. If you raise the\neffort to recover quality, the cost per request rises with it, and`max_completion_tokens` caps\nreasoning and visible output together.\n- +**The upgrade buys six months.**`gpt-5.1` 2025-11-13 itself retires on Azure on July 13, 2027,\nand OpenAI's API already lists`gpt-5.1` as deprecated, with an April 1, 2027 shutdown. A\ndeployment that rides the January 14 upgrade will face another migration by July.\n\nThat last point is the reason not to treat the automatic upgrade as the migration. It keeps the endpoint up. It doesn't test the new model on your traffic, and it lands you on a version with its own date.\n\nThe April 14 retirement of `gpt-4o` 2024-08-06 and 2024-11-20 also lists `gpt-5.1` as the\nreplacement. For `gpt-4o-mini` and the GPT-4.1 models, no replacement is listed yet, so nobody can\ntell you today which model a Standard deployment of them will be upgraded to.\n\n## Azure OpenAI and OpenAI retire the same models on different dates\n\nAzure and OpenAI each run their own schedule for the same model families:\n\n| Model and version | OpenAI API | Azure OpenAI | \n|---|---|---|\n| `gpt-4o` 2024-05-13 | Shut down October 23, 2026 | January 14, 2027 | \n| `gpt-4o` 2024-08-06 and 2024-11-20,`gpt-4o-mini` ,`gpt-4.1` ,`gpt-4.1-mini` | January 14, 2027 | April 14, 2027 | \n| `gpt-4.1-nano` | Shut down October 23, 2026 | April 14, 2027 | \n| `o3` ,`o3-pro` | December 11, 2026 | November 19, 2026 | \n| `o1` ,`o3-mini` ,`o4-mini` | Shut down October 23, 2026 | November 19, 2026 | \n| `gpt-5` ,`gpt-5-nano` | December 11, 2026 | February 9, 2027 | \n| `gpt-5-mini` | December 11, 2026 | July 13, 2027 | \n| `gpt-5.1` | April 1, 2027 | July 13, 2027 | \n\nTwo consequences follow. A team on both platforms can't plan from one page, because a fallback\nfrom Azure to OpenAI (or the reverse) can point at a model that is already gone on the other\nside. And the replacements differ: OpenAI's listed replacement for `gpt-4o` and `gpt-4.1` is\n`gpt-6-luna`, while Azure's is `gpt-5.1`. The\n[GPT-4o and GPT-4.1 shutdown guide](https://www.evalshift.dev/blog/gpt-4o-gpt-4-1-deprecation) covers OpenAI's side.\n\n## Fine-tuned models outlive their base model on Azure\n\nOn Azure, a fine-tuned model retires in two phases: training, when you can no longer create new fine-tunes of that base, and deployment, when existing fine-tuned deployments stop serving. The deployment dates run months past the base model's retirement:\n\n| Base model and version | Base model retires | Training retires (existing customers) | Fine-tuned deployments retire | \n|---|---|---|---|\n| `gpt-4o` 2024-08-06 | April 14, 2027 | No earlier than April 1, 2027 | October 1, 2027 | \n| `gpt-4o-mini` 2024-07-18 | April 14, 2027 | No earlier than April 1, 2027 | October 1, 2027 | \n| `gpt-4.1` 2025-04-14 | April 14, 2027 | No earlier than April 14, 2027 | October 14, 2027 | \n| `gpt-4.1-mini` 2025-04-14 | April 14, 2027 | No earlier than April 14, 2027 | October 14, 2027 | \n| `gpt-4.1-nano` 2025-04-14 | April 14, 2027 | No earlier than April 14, 2027 | April 13, 2028 | \n| `o4-mini` 2025-04-16 | November 19, 2026 | No earlier than May 19, 2027 | November 19, 2027 | \n\nFor customers who haven't fine-tuned that base before, training retires with the base model. The\n`o4-mini` row is the one to double-check: the retirement schedule says November 19, 2027, but the\nlifecycle policy page, last updated July 24, 2026, still says October 16, 2027 for deployments and\nApril 16, 2027 rather than May 19 for training.\n\nThis is the biggest difference from OpenAI's API, where fine-tunes of `gpt-4o` and `gpt-4.1` stop\nserving on January 14, 2027 together with their base. If you run fine-tunes on both platforms, the\nAzure copy is the one you can still record a baseline from in mid-2027.\n\n## How Azure's model lifecycle works\n\nMicrosoft's lifecycle policy gives every model one of five stages:\n\n| Stage | New deployments | Existing deployments | \n|---|---|---|\n| Preview | Yes | Work, but weights and API can change, and the model can be force-upgraded or retired with 30 days' notice | \n| Generally available (GA) | Yes | Work | \n| Legacy (optional) | Yes, until deprecation | Work | \n| Deprecated | Only subscriptions that have deployed this exact model version | Work | \n| Retired | No | Every request returns `410 Gone` | \n\nA GA model's retirement date is set when it launches, normally 18 months out, and Microsoft says\nthere is no separate announcement: the date is set at launch and readable from the Models API.\nNot every model gets 18 months (`gpt-6.1-terra` 2026-10-26 is listed to retire after 12), so read\nthe date rather than assume it. The standard pattern is deprecation at 12 months, the replacement available on Global Standard about\n90 days before retirement and on Provisioned about 30 days before, and retirement at 18 months.\nGA models from Anthropic, DeepSeek, Fireworks and Mistral AI on Foundry follow a 12-month\nlifecycle instead.\n\nThe Deprecated stage is where teams get caught. \"Existing customer\" is decided per Azure\nsubscription: whether that subscription has ever deployed that model version. A new subscription\nfor a new environment or region, even under the same tenant, can't deploy a Deprecated model. That\nis why `gpt-4o` 2024-11-20 matters: it is Legacy, not Deprecated, so it is the one `gpt-4o`\nversion a new subscription can deploy today, until Microsoft moves it to Deprecated. Microsoft\ncan also limit new customers in specific regions.\n\nMicrosoft notifies subscription owners with active deployments by email at least 60 days before a GA retirement and at least 30 days before a preview one, and posts Azure Service Health advisories. Service Health has no \"Microsoft Foundry\" service: filter on Azure OpenAI Service. Security or compliance problems can trigger an emergency retirement with shorter notice, and the policy FAQ says retirement dates can't be extended. Azure Government clouds usually run one version of a model at a time, with a 30-day overlap when a new one arrives.\n\n## Step 1: Inventory every deployment and its upgrade policy\n\nThe retirement schedule lists models, but what retires for you is deployments. List them per Azure OpenAI resource with the Azure CLI:\n\n```\naz cognitiveservices account deployment list \\\n  --resource-group my-rg --name my-openai-resource \\\n  --query \"[].{deployment:name, model:properties.model.name, version:properties.model.version, sku:sku.name, upgrade:properties.versionUpgradeOption}\" \\\n  --output table\n```\n\nAn empty `upgrade` column means the policy was never set, which behaves as\n`OnceCurrentVersionExpired`. Then check each model's lifecycle from the Models API, which is\nsubscription-scoped and per region:\n\n```\naz rest --method get \\\n  --url \"https://management.azure.com/subscriptions/$SUBSCRIPTION_ID/providers/Microsoft.CognitiveServices/locations/eastus2/models?api-version=2024-10-01\" \\\n  --query \"value[?model.name=='gpt-4o'].{version:model.version, status:model.lifecycleStatus, retires:model.deprecation.inference}\" \\\n  --output table\n```\n\nThe response is paged, so follow its `nextLink` if a model you deploy doesn't appear. The Models\nAPI names the stages differently from the docs and the portal. The API's\n`Deprecating` is what the docs call Deprecated (still serving, closed to new subscriptions), and\nthe API's `Deprecated` means Retired. Microsoft's own advice is to also compare\n`deprecation.inference` with today's date, because `lifecycleStatus` can lag. A check that alerts\non `Deprecated` alone fires after the model is already gone.\n\nDeployment names hide the model. A deployment called `chat-prod` tells you nothing, so record the\nmodel and version next to every deployment name your code, Terraform and eval harness use.\n\n## Step 2: Decide what each deployment should do on the date\n\nFor each Standard deployment on a retiring version, pick one of two paths:\n\n- +**Let it auto-upgrade, deliberately.** Fine for internal tools where an endpoint outage is worse\nthan a behavior change. Know the target (`gpt-5.1` for`gpt-4o` ) and its own July 13, 2027 date.\n- +**Migrate it yourself before the date.** For anything customer-facing, deploy the target model as\na new deployment, test it, and move traffic yourself. You choose the model, the`reasoning_effort` and the day.\n\nIf you take the second path, set `versionUpgradeOption` to `NoAutoUpgrade` on the old deployment\nonly if you'd rather have requests fail loudly than be silently answered by a model you didn't\ntest. Microsoft's docs note that the Azure CLI can show the setting but can't change it; use the\nportal, the REST API or Azure PowerShell.\n\nProvisioned deployments get no automatic path at all. Plan an in-place migration, which Microsoft says moves traffic over a 20 to 30 minute window without downtime, or a side-by-side one, and confirm you have PTU quota for the target model before the date.\n\n## Step 3: Record a baseline while the old version still answers\n\nAfter the retirement date, the old version returns `410 Gone`, and with an auto-upgraded deployment\nyou can't even tell from the endpoint when it changed. Record real production inputs now, 100 to\n300 per route, with the tools each one called, their arguments, the structured fields returned and\nthe refusals. That record is the only way to compare the new model against what your users actually\ngot. Log the `model` field of each response too: on Azure it normally carries the model name and\nversion, such as `o1-2024-12-17` in Microsoft's samples, which is how you see an upgrade land.\n\n## Step 4: Deploy the target side by side and replay real traffic\n\nCreate a new deployment of the candidate, `gpt-5.1` or anything newer that is available in your\nregion and deployment type, next to the old one. Replay the recorded inputs through both and\ncompare per route:\n\n- +tool selection and arguments, compared deterministically, not by eye;\n- +structured output fields and refusals;\n- +the request errors from parameters the new model rejects;\n- +cost per request and reasoning tokens, at each `reasoning_effort` you are considering;\n- +latency at the percentiles you alert on.\n\nFor a reasoning model, sample each case more than once: the answers vary from run to run, and a single pass can't tell a regression from noise.\n\n## Step 5: Move traffic and remove the old version from every fallback\n\nSwitch one route at a time behind a flag, starting with the lowest-risk one. Remove the retiring\ndeployment from every fallback list and retry chain before the date, because a fallback to a\nretired version fails exactly when you need it. Watch error rates, cost, latency and the response\n`model` field for a week after each switch.\n\n## Step 6: Put the next retirement date in CI\n\nEvery GA model on Azure has a retirement date from its first day, so the next migration is\nalready scheduled. Add a CI check that reads `deprecation.inference` from the Models API for every\nmodel version you deploy and fails when one is less than 90 days away. Set up an Azure Service\nHealth alert on Azure OpenAI Service as well, since the email goes to subscription owners, who are\noften not the engineers running the deployment.\n\n## Doing this with EvalShift\n\nDisclosure: I'm the founder of EvalShift, and steps 3 and 4 are the part it automates. The capture\nSDK's `wrap_openai` wraps any OpenAI client, including one pointed at your Azure resource's\n`/openai/v1/` endpoint, and records the deployment name you pass as `model`:\n\n``` python\nfrom openai import OpenAI\nfrom evalshift import capture\nfrom evalshift.adapters.openai import wrap_openai\n\nclient = wrap_openai(\n    OpenAI(base_url=\"https://my-openai-resource.openai.azure.com/openai/v1/\", api_key=AZURE_KEY)\n)\n\n@capture.agent(suite=\"support\", redact=True, tools=TOOLS)  # no-op unless EVALSHIFT_CAPTURE=1\ndef handle_ticket(message): ...\n```\n\n`compare` calls models through LiteLLM, which addresses an Azure deployment as\n`azure/<deployment-name>` and reads `AZURE_API_KEY`, `AZURE_API_BASE` and `AZURE_API_VERSION`:\n\n```\nevalshift capture sync\nevalshift compare --suite-name support --from azure/gpt-4o-prod --to azure/gpt-5-1-test\n```\n\nOne limit as of EvalShift 1.6.2: tool-call scoring maps a model ID to the OpenAI, Anthropic or\nGemini response shape by its name, and `azure/` IDs aren't mapped yet, so a suite that offers tools\nfails on them. Until that ships, run tool-calling suites against the same versions on OpenAI's API\nwhile both still serve there, for example `--from gpt-4o-2024-08-06 --to gpt-5.1-2025-11-13`, and\nsmoke-test the Azure deployment itself before the switch. EvalShift sends `temperature=0`; when a\nmodel rejects it, EvalShift resends without it and marks the run as non-deterministic in the\nreport, so set `defaults.samples_per_example` above 1 for the reasoning model. Docs:\n[/docs/sdk-adapters](https://www.evalshift.dev/docs/sdk-adapters), [/docs/captures](https://www.evalshift.dev/docs/captures),\n[/docs/migration-policy](https://www.evalshift.dev/docs/migration-policy).\n\n## FAQ\n\n**When does GPT-4o retire on Azure OpenAI?** `gpt-4o` version 2024-05-13 retires on January 14,\n2027, and versions 2024-08-06 and 2024-11-20 retire on April 14, 2027, together with `gpt-4o-mini`\n2024-07-18. Regional Standard deployments of 2024-05-13 already retired on March 31, 2026.\nFine-tuned `gpt-4o` and `gpt-4o-mini` deployments keep serving until October 1, 2027.\n\n**What happens to my Azure OpenAI deployment when its model retires?** Global Standard, Data Zone\nStandard and Standard deployments are upgraded under the same deployment name, to the listed\nreplacement where it's offered for that deployment type and region, unless their upgrade policy is `NoAutoUpgrade`, in which case they stop working. Provisioned\ndeployments are never upgraded automatically and stop serving. A retired model version returns\n`410 Gone` on every request.\n\n**What replaces GPT-4o on Azure OpenAI?** Microsoft lists `gpt-5.1` for every `gpt-4o` version,\npinned to version 2025-11-13 for the January 14 retirement. It is a reasoning model whose `reasoning_effort` defaults to `none` on Azure, and\nit retires on Azure on July 13, 2027. No replacement is listed yet for `gpt-4o-mini` or the GPT-4.1\nmodels, which Microsoft usually names 90 to 120 days before the retirement date.\n\n**Are Azure OpenAI retirement dates the same as OpenAI's?** No. Azure runs its own schedule. `gpt-4o`\n2024-08-06, `gpt-4o-mini`, `gpt-4.1` and `gpt-4.1-mini` shut down on OpenAI's API on January 14,\n2027 and on Azure on April 14, 2027, while `o3` and `o3-pro` were scheduled to retire on Azure on\nNovember 19, 2026, before OpenAI's December 11 date.\n\n**Can I get an extension on an Azure OpenAI model retirement?** No. Microsoft's lifecycle policy\nsays retirement dates aren't extendable. The only way to keep a workload running past the date is\nto move it to a supported model version, or to accept the automatic upgrade on a Standard\ndeployment.\n\n**Can a new Azure subscription still deploy gpt-4o?** Only the 2024-11-20 version, which is in the\nLegacy stage, until Microsoft moves it to Deprecated. Version 2024-08-06, `gpt-4o-mini` and the GPT-4.1 models are Deprecated, which means\nonly subscriptions that have already deployed that exact version can create new deployments of\nit, and a new subscription in the same tenant doesn't inherit that access.\n\n**How do I find out about Azure OpenAI model retirements?** Subscription owners with active\ndeployments get an email at least 60 days before a GA model retires and 30 days before a preview\nmodel does, and Azure Service Health shows advisories under Azure OpenAI Service. The Models API\nreturns each model's `lifecycleStatus` and `deprecation.inference` date, which a CI check can read.\n\n## Keep reading\n\n- +[LLM model deprecation schedule 2026](https://www.evalshift.dev/blog/llm-model-deprecation-schedule) — every OpenAI,\nAnthropic, Google, Azure and Bedrock shutdown date side by side.\n- +[GPT-4o and GPT-4.1 API shutdown on January 14, 2027](https://www.evalshift.dev/blog/gpt-4o-gpt-4-1-deprecation) — the\nOpenAI side of the same models, and the move to`gpt-6-luna` .\n- +[How to catch LLM regressions before a model upgrade ships](https://www.evalshift.dev/blog/llm-regression-testing-model-upgrade) — step 4 in full, with code and the failures it caught.\n- +[What actually breaks when you switch LLMs](https://www.evalshift.dev/blog/what-breaks-when-you-switch-llms) — why the\nworst regressions from a model swap read as correct.\n- +[LLM drift](https://www.evalshift.dev/blog/llm-drift) — how a model can change under a name you never edited.", "url": "https://wpnews.pro/news/azure-openai-model-retirement-every-date-auto-upgrades-and-how-to-migrate", "canonical_source": "https://www.evalshift.dev/blog/azure-openai-model-retirement", "published_at": "2026-11-19 16:46:09.316999+00:00", "updated_at": "2026-11-19 16:46:11.440561+00:00", "lang": "en", "topics": ["ai-policy", "large-language-models", "ai-infrastructure", "mlops"], "entities": ["Microsoft", "Azure OpenAI", "Microsoft Foundry", "gpt-4o", "gpt-5.1", "gpt-4.1", "o3", "o4-mini"], "also_reported_by": [], "takeaway": null, "recommendation": null, "alternates": {"html": "https://wpnews.pro/news/azure-openai-model-retirement-every-date-auto-upgrades-and-how-to-migrate", "markdown": "https://wpnews.pro/news/azure-openai-model-retirement-every-date-auto-upgrades-and-how-to-migrate.md", "text": "https://wpnews.pro/news/azure-openai-model-retirement-every-date-auto-upgrades-and-how-to-migrate.txt", "jsonld": "https://wpnews.pro/news/azure-openai-model-retirement-every-date-auto-upgrades-and-how-to-migrate.jsonld"}}