GPT-Live
The most recommended of 6 voice AI models
A real-time voice model that handles turn-taking, background noise and hands harder questions to a stronger model, praised for speed and natural voice. Critics often flag unnatural interjections like 'uh-huh'.
A real-time voice model that handles turn-taking, background noise and hands harder questions to a stronger model, praised for speed and natural voice. Critics often flag unnatural interjections like 'uh-huh'.
120 people gave an opinion
- recommend it
- 72
- criticize it
- 48
- approval share
- 60%
Recommended in 4 of 13 months
- November 2025: not recommended
- December 2025: not recommended
- January 2026: not recommended
- February 2026: not recommended
- March 2026: not recommended
- April 2026: not recommended
- May 2026: not recommended
- June 2026: not recommended
- July 2026: recommended
- August 2026: recommended
- September 2026: recommended
- October 2026: recommended
- November 2026: not recommended (month in progress)
Why it doesn’t carry the badge
- 72 people recommend itNeeds 10 or more, met.
- 1.5 recommenders for every criticNeeds 2 to 1 or better, not met.
- Recommended in 4 of 13 monthsNeeds 4 or more, met.
The same three rules apply to every product, and the badge can’t be bought. How the badge works
What people say about it
The reasons people give, with the number of comments that give each one. Open a reason to read what they said.
Why people recommend it
Each mark is one comment
Delegates hard questions to a stronger model
“I'm pretty sure that's a major part of this announcement? It delegates to GPT 5.5 which then uses tools.”
Comment from 8 Jul 2026 in “GPT‑Live” “What makes this one special is that it can delegate the task to a frontier model while talking to you.”
Comment from 8 Jul 2026 in “GPT‑Live” “The best feature is that it can delegate questions out to GPT-5.5 in the background”
Comment from 8 Jul 2026 in “GPT‑Live” - Comment from 8 Jul 2026 in “GPT‑Live”
Fast, natural voice responses
“I've been very impressed with response speed, intonation, and naturalness to the voice.”
Comment from 10 Jul 2026 in “Building a real-time AI tutor for 5-year-olds” “it's much faster, since it can reuse the background agent which now has a primed context”
Comment from 11 Sep 2026 in “GPT‑Live‑1 in the API” Better noise handling and turn-taking
“The new architecture makes sense, it seems many of the remaining problems like noise and interruptions are at the sound processing and integration level”
Comment from 8 Jul 2026 in “GPT‑Live” “it now has much better ability to understand the conversation and decide whether it should respond”
Comment from 8 Jul 2026 in “GPT‑Live” “GPT-Live is much better at ignoring background noise, including other people speaking.”
Comment from 8 Jul 2026 in “GPT‑Live” Useful for language practice and learning
“If it works as well as the demo, this already crushes most language learning apps.”
Comment from 8 Jul 2026 in “GPT‑Live” “This would be amazing for language listening/speaking practice.”
Comment from 8 Jul 2026 in “GPT‑Live” “extremely useful at refreshing my understanding of arbitrary science concepts in playful manner”
Comment from 21 Jul 2026 in “Ask HN: What's your experience with GPT-Live been like?” - Comment from 11 Sep 2026 in “GPT‑Live‑1 in the API”
What people warn about
Each mark is one comment
Random 'hmm' and 'uh-huh' interjections
“the interjected "uh-huh" or "yeah?" moments felt a little jarring”
Comment from 8 Jul 2026 in “GPT‑Live” “those "hmmm" "uh huh" interjected in the demo are pretty awful”
Comment from 8 Jul 2026 in “GPT‑Live” “Many times those little interjections don't feel natural.”
Comment from 8 Jul 2026 in “GPT‑Live” - Comment from 22 Jul 2026 in “Ask HN: What's your experience with GPT-Live been like?”
- Comment from 8 Jul 2026 in “GPT‑Live”
- Comment from 8 Jul 2026 in “GPT‑Live”
Voice sounds grating or cloying
“This voice is awful, possibly one of the worst AI/computer voices I have heard in like two years now”
Comment from 8 Jul 2026 in “GPT‑Live” “Can we have less terrible voices please? Nothing that sounds like a bubbly millennial.”
Comment from 8 Jul 2026 in “GPT‑Live” “I was hopeful that they avoided the well known sultry voice this go around, but alas”
Comment from 8 Jul 2026 in “GPT‑Live” Limited empathy and reasoning
“I am not convinced this model has enough empathy to satisfy most users on an emotional level.”
Comment from 8 Jul 2026 in “GPT‑Live” “GPT-Live isn't a particularly intelligent model. It needs to handoff for reasoning and more complex tasks.”
Comment from 21 Jul 2026 in “Ask HN: What's your experience with GPT-Live been like?” Mishears speech and wrong-language replies
“even seemingly GPT-Live-1 doesn't actually hear your pronunciation, it seems they all get passed transcripts”
Comment from 11 Sep 2026 in “GPT‑Live‑1 in the API” “I tried to talk to the openai version in (my bad) korean, and it responded in japanese”
Comment from 11 Sep 2026 in “GPT‑Live‑1 in the API”
Other voice AI models people recommend
The four most recommended other voice AI models, with the same counts for each, so you can compare like for like. The tick on each bar marks the 2 to 1 line the badge needs.
| Product | Recommend | Criticize | Approval | Months | Owned | Badge |
|---|---|---|---|---|---|---|
| GPT-LiveThis page | 72 | 48 | 60% | 4 | — | — |
| Gemini Live | 50 | 23 | 68% | 9 | — | Approved |
| Nvidia PersonaPlex 7B | 14 | 8 | 64% | 4 | — | — |
| gpt-realtime | 6 | 6 | 50% | 4 | — | — |
| Sesame voice demo | 5 | 1 | 83% | 3 | — | — |
What people said
40 of the 136 comments that mention it, newest first. Each links to the original comment.
“The regular model generally does not suffer the same issues he is demonstrating with the real time audio version.”
“I asked ChatGPT to stretch that part and make them stay on the road for longer, and it kinda did an OK job tbh”
“I tried to talk to the openai version in (my bad) korean, and it responded in japanese”
“it's much faster, since it can reuse the background agent which now has a primed context”
“this could be useful for the case of learning something by teaching it to someone else”
“even seemingly GPT-Live-1 doesn't actually hear your pronunciation, it seems they all get passed transcripts”
“This is the sort of thing that the GPT-Live model would excel at when it launches.”
“the new gpt-live feature tends to interrupt with 'human' sounding injections like "hmm", "huh" a bit too much for my liking”
“GPT-Live isn't a particularly intelligent model. It needs to handoff for reasoning and more complex tasks.”
“The model has done surprisingly well in conversations where it has to go back and forth across languages in the same response”
“extremely useful at refreshing my understanding of arbitrary science concepts in playful manner”
“I've been very impressed with response speed, intonation, and naturalness to the voice.”
“Super nice to be able to finally have voice mode search stuff in the background”
“it actually correctly counted and didn't leave out any number. I'm impressed”
“It gives me hope that, even if my disease progresses, I’ll be able to rely less on screens”
“I professionally use it to design pharmaceutical drugs, essentially I do my best to harness the good sides of it”
“It is astonishingly good with certain rare accents, and it responds well to longer input clips”
“We see this as the first version of a new generation; expect many improvements in the future.”
“Not super impressed by the model constantly interrupting the user in the other demos”
“If it works as well as the demo, this already crushes most language learning apps.”
“I think it sounds fine in my language, and it's an improvement over the previous one.”
“This voice is awful, possibly one of the worst AI/computer voices I have heard in like two years now”
“Can we have less terrible voices please? Nothing that sounds like a bubbly millennial.”
“the real value of the bot, which is to handle tasks via voice that an average human either would not or could not do”
“I'm pretty sure that's a major part of this announcement? It delegates to GPT 5.5 which then uses tools.”
“This would be amazing for language listening/speaking practice.”
“I am not convinced this model has enough empathy to satisfy most users on an emotional level.”
“It’s not quite lived up to my expectations but I think this gets much closer.”
“What makes this one special is that it can delegate the task to a frontier model while talking to you.”
“I was hopeful that they avoided the well known sultry voice this go around, but alas”
“The new architecture makes sense, it seems many of the remaining problems like noise and interruptions are at the sound processing and integration level”
“It's very good. I had one conversation that lasted a full hour while I was walking the dog”
“it now has much better ability to understand the conversation and decide whether it should respond”
“GPT-Live is much better at ignoring background noise, including other people speaking.”
“Many times those little interjections don't feel natural.”
“full-duplex architecture + delegation enables entirely new ways of human-AI interaction”
“The full duplex is awesome, and the feedback that it is getting what you're saying is ok, but in some of the demos was a little overkill.”
“those "hmmm" "uh huh" interjected in the demo are pretty awful”
“The best feature is that it can delegate questions out to GPT-5.5 in the background”
“the interjected "uh-huh" or "yeah?" moments felt a little jarring”
Counts come from 1,795,640 Hacker News comments, November 2025 to November 2026. How we count. Report a mistake.