<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>RADAR — Azure Foundry model retirement schedule</title>
    <link>https://radar.trulysovereign.ai</link>
    <atom:link href="https://radar.trulysovereign.ai/feeds/source/azure-foundry-model-retirement-schedule.xml" rel="self" type="application/rss+xml"/>
    <description>Changes RADAR detected at Azure Foundry model retirement schedule.</description>
    <language>en</language>
    <lastBuildDate>Thu, 24 Sep 2026 04:30:30 GMT</lastBuildDate>
    <item>
      <title>gpt-realtime-2 retirement date removed, gpt-realtime-2.1 added</title>
      <link>https://radar.trulysovereign.ai/items/f654bd7b-a520-4e76-973f-111f69f761f2</link>
      <guid isPermaLink="false">f654bd7b-a520-4e76-973f-111f69f761f2</guid>
      <pubDate>Wed, 23 Sep 2026 21:12:27 GMT</pubDate>
      <category>verify_only</category>
      <source url="https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule">Azure Foundry model retirement schedule</source>
      <description>The retirement date of 2026-08-31 was removed from gpt-realtime-2 (preview model launched 2026-05-06). A new model version gpt-realtime-2.1 was added to the schedule with a launch date of 2026-07-07 and no stated retirement date.

Why it matters: If you are using gpt-realtime-2 and planned migration work around the August 31 deadline, that deadline no longer exists. The model remains in preview with no published end-of-life, which means the retirement timeline is now undefined.

What to do: Check your deployment configs for gpt-realtime-2 references. If you scheduled migration work for before August 31, deprioritize it until a new retirement date is published or evaluate whether to migrate to gpt-realtime-2.1 based on your release criteria for preview models.</description>
    </item>
    <item>
      <title>gpt-realtime-2.1 and mini moved to GA on 2026-07-07</title>
      <link>https://radar.trulysovereign.ai/items/68b7879f-1277-4f33-a8f6-6843ca33d56e</link>
      <guid isPermaLink="false">68b7879f-1277-4f33-a8f6-6843ca33d56e</guid>
      <pubDate>Mon, 21 Sep 2026 23:43:42 GMT</pubDate>
      <category>verify_only</category>
      <source url="https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule">Azure Foundry model retirement schedule</source>
      <description>gpt-realtime-2.1 and gpt-realtime-2.1-mini changed status from Preview to GA effective 2026-07-07. Both models have a retirement date of 2027-06-25.

Why it matters: GA status means these models now carry production SLA guarantees and support commitments that Preview versions do not. Any production workloads currently using these models are now covered under standard Azure support terms.

What to do: If you are using either model in production, verify your deployment is receiving GA-level support and SLA coverage. If you were blocking production use pending GA status, that gate is now cleared.</description>
    </item>
    <item>
      <title>gpt-4.1-nano retirement extended to 2027-04-14</title>
      <link>https://radar.trulysovereign.ai/items/da960a09-8d62-4cd9-aa80-fc01b818f5fc</link>
      <guid isPermaLink="false">da960a09-8d62-4cd9-aa80-fc01b818f5fc</guid>
      <pubDate>Mon, 21 Sep 2026 23:43:23 GMT</pubDate>
      <category>migrate</category>
      <source url="https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule">Azure Foundry model retirement schedule</source>
      <description>The retirement date for gpt-4.1-nano moved from 2026-10-14 to 2027-04-14, a six-month extension. The model entered Legacy status on 2025-04-14.

Why it matters: If you are running gpt-4.1-nano in production, you now have until mid-April 2027 to complete migration instead of mid-October 2026. This changes the timeline for any scheduled replacement work.

What to do: Check your deployment manifests and infrastructure-as-code for references to gpt-4.1-nano. If found, update your migration plan to reflect the new 2027-04-14 deadline and communicate the revised timeline to stakeholders.</description>
    </item>
    <item>
      <title>Azure retires gpt-4.1-nano and gpt-image-1.5 six months earlier than announced</title>
      <link>https://radar.trulysovereign.ai/items/1025bd98-ec2e-4bba-a135-160375cd0aa8</link>
      <guid isPermaLink="false">1025bd98-ec2e-4bba-a135-160375cd0aa8</guid>
      <pubDate>Mon, 21 Sep 2026 23:42:57 GMT</pubDate>
      <category>migrate</category>
      <source url="https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule">Azure Foundry model retirement schedule</source>
      <description>gpt-4.1-nano retirement moved from 2027-04-14 to 2026-10-14. gpt-image-1.5 retirement moved from 2027-06-16 to 2026-12-16. Both dates accelerated by six months.

Why it matters: Any deployments using these models will stop working on the new dates. Six months less migration runway means work scheduled for Q2 2027 must now complete by Q4 2026.

What to do: Audit Azure deployments for gpt-4.1-nano and gpt-image-1.5. If found, schedule migration to successor models before 2026-10-14 and 2026-12-16 respectively.</description>
    </item>
    <item>
      <title>Azure o1/o3 family retirement consolidated to 2026-11-19</title>
      <link>https://radar.trulysovereign.ai/items/23bc0de3-09e4-4b40-8fef-f1254dbbec6d</link>
      <guid isPermaLink="false">23bc0de3-09e4-4b40-8fef-f1254dbbec6d</guid>
      <pubDate>Sat, 12 Sep 2026 03:31:37 GMT</pubDate>
      <category>migrate</category>
      <source url="https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule">Azure Foundry model retirement schedule</source>
      <description>o1-pro, o3, o3-deep-research, and o3-pro moved from GA to Deprecated status. All o1/o3 models now retire 2026-11-19 (previously scattered between 2026-10-01 and 2026-12-26). o3-mini and o4-mini successor changed from o4-mini to gpt-5.6-terra. Fine-tuning retirement for o4-mini shifted from 2027-04-16 to 2027-05-19.

Why it matters: If you are running o1-pro, o3, o3-deep-research, or o3-pro in production, they are now deprecated with 9 weeks until retirement. The consolidated sunset date means you cannot stagger migrations. The successor model change means existing migration plans targeting o4-mini must be rewritten to target gpt-5.6-terra or gpt-5.6-sol.

What to do: Grep your codebase for model identifiers o1-pro, o3, o3-deep-research, o3-pro, o3-mini, o4-mini. If any are in use, schedule migration to gpt-5.6-sol (for o1/o3 reasoning models) or gpt-5.6-terra (for o3-mini/o4-mini) before 2026-11-19. If you have fine-tuned o4-mini models, note the new training retirement date of 2027-05-19.</description>
    </item>
    <item>
      <title>Two image models added to retirement schedule, sunset 9 September 2027</title>
      <link>https://radar.trulysovereign.ai/items/a6e8fa9f-1df0-4980-b269-29cd94f20820</link>
      <guid isPermaLink="false">a6e8fa9f-1df0-4980-b269-29cd94f20820</guid>
      <pubDate>Thu, 10 Sep 2026 20:30:21 GMT</pubDate>
      <category>verify_only</category>
      <source url="https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule">Azure Foundry model retirement schedule</source>
      <description>gpt-image-2.5-flare and gpt-image-2.5-sunburst reached GA on 9 September 2026 and are scheduled for retirement on 9 September 2027. Both models now appear on the Azure Foundry model retirement schedule.

Why it matters: If you adopt either model after GA, you have exactly one year before forced migration. The retirement date is fixed at announcement, which is unusually short for a GA model lifecycle.

What to do: Check your image generation endpoints for references to gpt-image-2.5-flare or gpt-image-2.5-sunburst. If you are evaluating these models, factor the 9 September 2027 cutoff into your adoption decision now.</description>
    </item>
    <item>
      <title>Claude Fable 5.1 retires 5 December 2027</title>
      <link>https://radar.trulysovereign.ai/items/c5855813-268c-4213-8f94-7cf3f38ef45c</link>
      <guid isPermaLink="false">c5855813-268c-4213-8f94-7cf3f38ef45c</guid>
      <pubDate>Thu, 03 Sep 2026 03:20:06 GMT</pubDate>
      <category>migrate</category>
      <source url="https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule">Azure Foundry model retirement schedule</source>
      <description>Azure Foundry added a retirement date of 2027-12-05 for the claude-fable-5-1 model. The model is currently in Preview status.

Why it matters: Any code or configuration using claude-fable-5-1 by name will stop working after that date. Preview models typically have shorter lifespans than GA releases, and 15 months is the notice window.

What to do: Grep your codebase and infrastructure-as-code for 'claude-fable-5-1' references. If found, add a calendar reminder for Q3 2027 to migrate to whatever replaces it, or pin to claude-fable-5 if that model remains supported.</description>
    </item>
    <item>
      <title>claude-opus-4-1 retired August 5, 2026; upgrade path to claude-opus-5</title>
      <link>https://radar.trulysovereign.ai/items/cf2722ee-e40f-454a-bc92-5775238664a9</link>
      <guid isPermaLink="false">cf2722ee-e40f-454a-bc92-5775238664a9</guid>
      <pubDate>Tue, 01 Sep 2026 20:40:33 GMT</pubDate>
      <category>migrate</category>
      <source url="https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule">Azure Foundry model retirement schedule</source>
      <description>claude-opus-4-1 marked Retired with effective date 2026-08-05. claude-opus-5 added to supported models list. Two new preview models added: claude-fable-5-1 and claude-mythos-5-1.

Why it matters: Any code or configuration explicitly requesting claude-opus-4-1 will fail after August 5, 2026. If you are using version pinning or explicit model identifiers rather than latest-stable aliases, calls will return errors once the retirement date passes.

What to do: Grep your codebase and infrastructure-as-code for 'claude-opus-4-1' and replace with 'claude-opus-5'. Test the replacement in staging before August 5, 2026.</description>
    </item>
    <item>
      <title>Three MAI-Image-2.5 variants added to retirement schedule, sunset October 2026</title>
      <link>https://radar.trulysovereign.ai/items/c00829cd-4938-43d2-9f3f-f7eaa759fb5b</link>
      <guid isPermaLink="false">c00829cd-4938-43d2-9f3f-f7eaa759fb5b</guid>
      <pubDate>Sat, 29 Aug 2026 05:59:09 GMT</pubDate>
      <category>migrate</category>
      <source url="https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule">Azure Foundry model retirement schedule</source>
      <description>MAI-Image-2.5, MAI-Image-2.5-Flash, and MAI-Image-2.5-Pro added to Azure Foundry retirement schedule. All three entered Preview on June 2 or June 19, 2026, and will be retired October 1, 2026. MAI-Image-2e status changed from Preview to Retired.

Why it matters: If you are using or planning to use any MAI-Image-2.5 variant, you have a four-month window from preview to retirement. Code written against these models will break on October 1, 2026 unless migrated to a successor model.

What to do: Grep your codebase for 'MAI-Image-2.5', 'MAI-Image-2.5-Flash', and 'MAI-Image-2.5-Pro'. If found, schedule migration work to complete before October 1, 2026. If not using these models, no action required.</description>
    </item>
    <item>
      <title>Azure Foundry splits Claude models into two hosting versions with new retirement dates</title>
      <link>https://radar.trulysovereign.ai/items/f0d559ef-a8a6-4d04-adbd-3a1e98096d5d</link>
      <guid isPermaLink="false">f0d559ef-a8a6-4d04-adbd-3a1e98096d5d</guid>
      <pubDate>Fri, 28 Aug 2026 21:46:44 GMT</pubDate>
      <category>verify_only</category>
      <source url="https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule">Azure Foundry model retirement schedule</source>
      <description>Azure Foundry now offers Claude models in two versions: Version 1 (Anthropic infrastructure) and Version 2 (Azure infrastructure). Multiple models moved from Preview to GA status. claude-opus-4-1 was removed from the schedule (previously marked Retired 2026-08-05). New model claude-fable-5 added with retirement date 2027-12-05. claude-haiku-4-5 Preview variant removed; GA variants now exist for both Version 1 and Version 2, both retiring 2026-10-19.

Why it matters: If you are calling Claude models by name without specifying version, your requests may now route to different infrastructure (Anthropic vs Azure) with potentially different latency, availability, or billing characteristics. The removal of claude-opus-4-1 from the schedule means any code still referencing it is calling a model that no longer appears in the official retirement timeline.

What to do: Audit your model identifiers to confirm which version (1 or 2) you are using. If you are calling claude-opus-4-1, verify whether it still responds or migrate to claude-opus-5 (GA, retires 2027-07-08). If you rely on claude-haiku-4-5, confirm whether you are on Version 1 or 2 and note the 2026-10-19 retirement date for both.</description>
    </item>
    <item>
      <title>Sora-2 preview retirement delayed one month to 15 October 2026</title>
      <link>https://radar.trulysovereign.ai/items/45db7e68-2aed-4776-bb75-8dc35012c6f1</link>
      <guid isPermaLink="false">45db7e68-2aed-4776-bb75-8dc35012c6f1</guid>
      <pubDate>Tue, 25 Aug 2026 06:49:24 GMT</pubDate>
      <category>verify_only</category>
      <source url="https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule">Azure Foundry model retirement schedule</source>
      <description>The retirement date for the sora-2 preview model moved from 15 September 2026 to 15 October 2026.

Why it matters: If you planned migration work around the September deadline, you now have four additional weeks. If you already migrated, the extension changes nothing.

What to do: Check your migration schedule for sora-2. If you have not started and the September date was tight, you can defer by up to four weeks. If you already migrated or the date was not a constraint, no action needed.</description>
    </item>
    <item>
      <title>gpt-4.1-nano retirement extended six months to 2027-04-14</title>
      <link>https://radar.trulysovereign.ai/items/ebad5c37-05e6-4d61-96a7-661d9006c8d5</link>
      <guid isPermaLink="false">ebad5c37-05e6-4d61-96a7-661d9006c8d5</guid>
      <pubDate>Thu, 20 Aug 2026 01:15:58 GMT</pubDate>
      <category>migrate</category>
      <source url="https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule">Azure Foundry model retirement schedule</source>
      <description>Azure Foundry moved gpt-4.1-nano retirement from 2026-10-14 to 2027-04-14. The model's status label changed from 'Deprecated' to 'Legacy' on 2025-04-14. Three other models (gpt-4.1, gpt-4.1-mini, gpt-4o) also changed to 'Legacy' status with 2027-04-14 retirement dates.

Why it matters: If you planned migration work around the October 2026 deadline, you now have six additional months. Teams that deprioritized gpt-4.1-nano migration because of the tight timeline can reschedule without urgency.

What to do: Check your migration backlog for gpt-4.1-nano work scheduled before 2027-04-14. If you deferred it due to the October deadline, update the target date and reprioritize against other work.</description>
    </item>
    <item>
      <title>DeepSeek-V4-Flash-0731 preview ends 2026-12-03, GA 2026-07-31</title>
      <link>https://radar.trulysovereign.ai/items/2a3739cf-3ab9-49a7-ab94-0249afd03331</link>
      <guid isPermaLink="false">2a3739cf-3ab9-49a7-ab94-0249afd03331</guid>
      <pubDate>Wed, 19 Aug 2026 18:36:14 GMT</pubDate>
      <category>migrate</category>
      <source url="https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule">Azure Foundry model retirement schedule</source>
      <description>Azure added DeepSeek-V4-Flash-0731 to the retirement schedule. The model enters GA on 2026-07-31 and preview access ends on 2026-12-03.

Why it matters: If you are testing this model in preview, you have until 2026-12-03 to migrate to the GA version or switch models. Preview endpoints will stop working after that date.

What to do: Check if any code or configs reference DeepSeek-V4-Flash-0731 in preview. If so, schedule migration to the GA version before 2026-12-03.</description>
    </item>
    <item>
      <title>Four Azure Foundry models retired, MAI-Transcribe-1 retiring 15 September 2026</title>
      <link>https://radar.trulysovereign.ai/items/1a010efc-3b68-4ca7-909c-3aa558516e6b</link>
      <guid isPermaLink="false">1a010efc-3b68-4ca7-909c-3aa558516e6b</guid>
      <pubDate>Tue, 18 Aug 2026 01:15:03 GMT</pubDate>
      <category>migrate</category>
      <source url="https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-retirement-schedule">Azure Foundry model retirement schedule</source>
      <description>DeepSeek-R1, MAI-Image-2, Stable-Diffusion-3.5-Large, Stable-Image-Core, and Stable-Image-Ultra moved from Legacy/Preview/Deprecated to Retired status between 31 July and 15 August 2026. MAI-Transcribe-1 added to retirement schedule with cutoff date 15 September 2026. MAI-Transcribe-1.5 added as successor model.

Why it matters: Retired models are no longer callable via API. Any code still referencing DeepSeek-R1, MAI-Image-2, or the three Stable models will fail. MAI-Transcribe-1 stops working in 28 days.

What to do: Grep your codebase for 'DeepSeek-R1', 'MAI-Image-2', 'Stable-Diffusion-3.5-Large', 'Stable-Image-Core', 'Stable-Image-Ultra', and 'MAI-Transcribe-1'. Remove or replace references before 15 September 2026 for MAI-Transcribe-1; the other five are already dead.</description>
    </item>
  </channel>
</rss>
