{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreib5nhqtmxp6dwapyvqrrj72shpefl6ihhpy46zu5nbju3ws6k4l5y",
    "uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mljhzanyffh2"
  },
  "path": "/t/realtime-regression-in-non-english-production-voice-agents-gpt-realtime-mini-vs-gpt-realtime-mini-2025-10-06/1380643#post_3",
  "publishedAt": "2026-05-10T18:49:02.000Z",
  "site": "https://community.openai.com",
  "tags": [
    "Deprecations | OpenAI API"
  ],
  "textContent": "vb:\n\n> Hi!\n>\n> Thank you for raising this.\n>\n> You are probably already aware that you can still use gpt-realtime-mini-2025-10-06. You do not need to use the undated model slug, which currently points to the December snapshot.\n>\n> For production systems, dated model versions are usually the better choice because they help keep behavior more consistent.\n>\n> Since I do not see a deprecation notice for the older snapshot, using it seems reasonable here.\n>\n> In this case, I would suggest staying on the older snapshot for now and waiting for a possible gpt-realtime-mini-2 release, or reviewing the Realtime model best practices here:\n\nThank you for the reply.\n\nJust to clarify one important point: `gpt-realtime-mini-2025-10-06` does appear to be listed on the official OpenAI deprecations page Deprecations | OpenAI API\n\nOn the current deprecations page, under the section for legacy GPT model snapshots, I see the following entry:\n\nShutdown date: 2026-07-23\nModel snapshot: `gpt-realtime-mini-2025-10-06`\nSubstitute model: `gpt-realtime-mini`\n\nSo while we can still use the dated snapshot for now, the problem is that it appears to have a scheduled shutdown date, and the listed substitute model does not currently behave equivalently in our non-English/Romanian production voice-agent flows.\n\nThat is the core issue we are trying to surface: the dated snapshot is currently production-stable for us, but the documented replacement introduces worse language quality and more hallucination/confabulation against supplied business data.\n\nFor that reason, we are looking for either:\n\n  1. confirmation of whether this deprecation entry is accurate;\n\n  2. a migration path for production users affected by language-specific regressions;\n\n  3. or guidance on how to escalate this as a production-impacting model quality regression before the shutdown date.\n\n\n\n\nWe are happy to provide side-by-side transcripts showing the behavioral difference.",
  "title": "Realtime regression in non-English production voice agents: gpt-realtime-mini vs gpt-realtime-mini-2025-10-06"
}