{
"$type": "site.standard.document",
"bskyPostRef": {
"cid": "bafyreia62dceqdglwtlfc34ctiulijwgez4jzv3cujidw4skwez7m2qqii",
"uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mlc4rbnze422"
},
"path": "/t/new-realtime-voice-models-in-the-api/1380471#post_4",
"publishedAt": "2026-05-07T20:52:38.000Z",
"site": "https://community.openai.com",
"textContent": "I already did a quick test with gpt-realtime-2 using my Cheeky-Razor persona.\n\nThe audio output felt more natural than with previous models: clearer, better paced and easier to follow.\n\nI tested it in English and as a non-native English speaker, I found it easy to understand without having to concentrate too hard, which is actually a very useful improvement​",
"title": "New Realtime Voice Models in the API"
}