AI Wiki

Explore AI Wiki

Citation and evidence

Advanced Voice Mode

20 min full readUpdated 26 references

This article's verification

Report a problem with this article

More

Use this article

Raw MarkdownExplore connections

Improve this page

Suggest editRevision historyDiscussion

Browse categories

ChatGPTOpenAISpeech & Audio AIVoice AI

Cite this article

Advanced Voice Mode is a real-time, spoken conversation feature in ChatGPT, developed by OpenAI, that lets users talk with the assistant using natural speech and receive spoken replies. It is built on the native, end-to-end audio capabilities of the GPT-4o model, which processes speech directly rather than first transcribing it to text. OpenAI first demonstrated the capability at the GPT-4o launch on May 13, 2024, began an alpha release to a small group of ChatGPT Plus users on July 30, 2024, and expanded it to all Plus and Team subscribers beginning September 24, 2024.[1][2][3] OpenAI reported that GPT-4o can respond to audio inputs in as little as 232 milliseconds, with an average of 320 milliseconds, which it described as similar to human response time in a conversation, versus averages of 2.8 seconds and 5.4 seconds for the earlier GPT-3.5 and GPT-4 voice pipelines.[1] The feature is distinct from the GPT-4o model itself, from the developer-facing Realtime API, and from OpenAI's Voice Engine text-to-speech model.

OpenAI now calls the overall product ChatGPT Voice. On July 8, 2026, it made a new full-duplex model family, GPT-Live, the default engine for ChatGPT Voice, and as of September 2026 OpenAI's help center lists three voice options: Live (GPT-Live-1 or GPT-Live-1 mini), Advanced (which OpenAI describes as the previous real-time Voice experience, still used for video and screen sharing on mobile), and Standard.[20][22] On September 23, 2026, OpenAI added plugin support and voice access to ChatGPT Work on web and mobile, and its announcement post said Voice could be powered by GPT-6 Astra, Sol and Luna (Astra had been available in Voice since September 9).[7][23]

Overview

Advanced Voice Mode is the consumer-facing ChatGPT experience that turns spoken audio into a fluid, two-way conversation with the assistant. It is the successor to ChatGPT's earlier voice feature and is defined by three traits: very low latency, the ability to perceive and express tone and emotion, and the ability to be interrupted mid-sentence so the conversation feels like talking to a person rather than dictating to a transcription tool.

OpenAI announced spoken conversation for ChatGPT on September 25, 2023, with the voice feature rolling out over the following two weeks to Plus and Enterprise subscribers as an opt-in beta in the iOS and Android apps.[4][7] OpenAI made voice available to all free users on November 21, 2023.[7] That original system, now referred to as Standard Voice Mode, worked by chaining three separate models together: OpenAI's open-source Whisper model transcribed the user's speech to text, GPT-3.5 or GPT-4 generated a written reply, and a separate text-to-speech model read that reply aloud.[1][4] Because each step ran sequentially, the pipeline introduced noticeable latency, averaging 2.8 seconds with GPT-3.5 and 5.4 seconds with GPT-4, and the main model lost information such as tone, multiple speakers, and background noise along the way.[1]

Advanced Voice Mode replaced that pipeline with a single model. With GPT-4o, OpenAI trained one neural network end-to-end across text, vision, and audio, so that audio input and audio output are handled by the same model without an intermediate text-conversion step.[1] This allows the model to perceive characteristics of speech that a transcript cannot capture, and to respond with a spoken voice that can vary in tone and pacing.

Naming

OpenAI's own wording has shifted over time. The 2023 feature was introduced as "Voice (Beta)" and called Voice Mode; the GPT-4o version launched in 2024 as "Advanced Voice Mode," and OpenAI's release notes described the older version as "Standard Voice Mode."[1][7] From August 2025 onward, OpenAI's release notes increasingly used the name "ChatGPT Voice" for the product as a whole.[7] OpenAI's July 2026 GPT-Live announcement refers to "ChatGPT Advanced Voice Mode" as an earlier, turn-based generation, and the current help center article is titled "ChatGPT Voice," with Advanced listed as one of three selectable options.[20][22]

Technical design

The defining technical characteristic of Advanced Voice Mode is that it relies on GPT-4o's native speech-to-speech processing. Rather than converting audio to text, reasoning over the text, and synthesizing new audio with three different systems, GPT-4o accepts audio directly and emits audio directly. OpenAI stated that all inputs and outputs are processed by the same neural network, and contrasted this with the older pipeline, in which the main model "can't directly observe tone, multiple speakers, or background noises, and it can't output laughter, singing, or express emotion."[1]

This single-model design produces two practical effects. First, it reduces latency dramatically compared with the chained Standard Voice Mode, enabling near-instant, back-and-forth exchanges in which the user can interrupt the assistant mid-sentence and have it stop and respond. Second, it preserves and can generate paralinguistic information: OpenAI said GPT-4o can sense emotional intonation in a user's voice, such as sadness or excitement, and the model can produce expressive speech that the older pipeline could not.[1][2] These behaviors are properties of the underlying GPT-4o model; Advanced Voice Mode is the consumer-facing ChatGPT feature that exposes them in a live conversational interface.

Advanced Voice Mode still worked in discrete turns. In its July 2026 GPT-Live announcement, OpenAI described turn-based voice models such as Advanced Voice Mode as having to wait for the user to stop speaking before responding, and said that because turn detection is based on silence, "even a brief pause or background noise could be mistaken for the end of turn."[20] GPT-Live replaced this with a full-duplex design that continuously processes input while generating output, and it delegates questions that need search, reasoning or agentic work to a separate frontier model while the conversation continues.[20]

GenerationOpenAI's descriptionIntroduced in ChatGPT
Standard Voice ModeCascade of speech-to-text model, language model (GPT-3.5 or GPT-4 at launch) and text-to-speech model[1][20]September 2023
Advanced Voice ModeSingle GPT-4o model handling audio in and out; turn-based[1][20]July 2024 (alpha); September 2024 (broad rollout)
GPT-Live (Live option)Full-duplex model that listens and speaks at the same time, delegating harder work to a frontier model[20]July 8, 2026 (default from launch)

Expanded article table

Relationship to the Realtime API and Voice Engine

It is worth distinguishing Advanced Voice Mode from related OpenAI systems. The Realtime API, which OpenAI introduced as a public beta for paid developers on October 1, 2024, gives programmatic access to natural, low-latency speech-to-speech conversations that OpenAI compared to ChatGPT's Advanced Voice Mode.[17] OpenAI made the Realtime API generally available on August 28, 2025, alongside a new speech-to-speech model, gpt-realtime, and made GPT-Live-1 generally available in the API on September 10, 2026.[18][19] Voice Engine is a text-to-speech model that OpenAI previewed on March 29, 2024; it generates speech resembling a speaker from text plus a single 15-second audio sample. OpenAI said it first developed Voice Engine in late 2022 and has used it to power the preset voices in its text-to-speech API as well as "ChatGPT Voice and Read Aloud."[16] Advanced Voice Mode specifically refers to the in-app ChatGPT experience built on GPT-4o.

Features

Advanced Voice Mode is designed around natural, free-flowing conversation. Its principal features include:

  • Real-time, interruptible dialogue, allowing users to break in while the assistant is speaking.
  • Perception of tone and emotion in the user's speech, and the ability to convey emotion and varied intonation in its own responses.[1][2]
  • Multiple preset voices. At the alpha launch, four voices were available: Juniper, Breeze, Cove, and Ember.[2] The September 2024 general release added five more, Arbor, Maple, Sol, Spruce, and Vale, bringing the total to nine, alongside improved accents.[3] As of September 2026, OpenAI's help center lists the same nine voices, plus two additional voices, Viola and Rio, that users in Brazil may see.[22]
  • Support for Custom Instructions and Memory, so the assistant can apply user preferences and recall details from earlier conversations.[3]
  • Live video and screen sharing, added on December 12, 2024, which let users point a phone camera at objects or share their screen and ask ChatGPT about what it sees in real time.[5][7]
  • Language translation, added for paid users on June 7, 2025: users can ask Voice to translate between languages and it continues translating until told to stop.[7]

A version of Advanced Voice powered by GPT-4o mini later reached free users; OpenAI began rolling out a daily preview to all free ChatGPT users on February 25, 2025, saying its conversation pace and tone were "similar to the GPT-4o version while being more cost effective to serve."[6] As of November 2025, OpenAI's help center said subscribers' voice sessions began with GPT-4o and continued on GPT-4o mini once the day's GPT-4o minutes were used up.[12]

Release history

DateMilestone
September 25, 2023Voice (Whisper plus text-to-speech) announced; rollout to Plus and Enterprise users on iOS and Android begins[4][7]
November 21, 2023Voice made available to all free users[7]
May 13, 2024GPT-4o launched; native voice capability demonstrated at OpenAI's Spring Update[1]
May 19, 2024OpenAI pauses the "Sky" voice[8][13]
June 2024OpenAI delays Advanced Voice Mode to improve its safety measures[2]
July 30, 2024Advanced Voice Mode alpha released to a small group of ChatGPT Plus users[2]
September 24, 2024Broad rollout begins to all Plus and Team users, with new voices and a new look[3]
October 1, 2024Realtime API public beta brings similar speech-to-speech conversations to developers[17]
October 22, 2024Advanced Voice becomes available to Plus users in the EU, Switzerland, Iceland, Norway, and Liechtenstein[15]
November 19, 2024Advanced Voice begins rolling out on chatgpt.com (web)[7]
December 12, 2024Live video, screen sharing and image upload added in mobile apps during the "12 Days of OpenAI"; Santa voice added[5][7]
February 25, 2025A GPT-4o mini-powered version begins rolling out to free users as a daily preview[6]
June 7, 2025Upgrade for paid users with more natural intonation and live translation[7]
July 17, 2025The June upgrades roll out to free users[7]
August 7, 2025Voice works with custom GPTs; near-unlimited use for Plus and hours per day for free users; Standard Voice Mode retirement announced for 30 days later[7]
September 9, 2025OpenAI keeps Standard Voice available after user feedback[7]
November 25, 2025Voice integrated directly into the chat thread, with the separate full-screen interface kept as an option[7][10][11]
January 15, 2026Voice retired from the ChatGPT macOS app (announced December 11, 2025)[7]
July 8, 2026GPT-Live-1 and GPT-Live-1 mini become the default models for ChatGPT Voice[7][20]
July 23, 2026ChatGPT Voice available in Work and Codex in the ChatGPT desktop app[7]
September 9, 2026Voice can use GPT-5.6 or GPT-6 Astra for harder questions; new daily limits[7]
September 23, 2026Plugins in Voice and Voice in ChatGPT Work on web and mobile[7]; OpenAI's post adds GPT-6 Sol and Luna alongside Astra as models that can power Voice[23]

Expanded article table

When the general rollout began on September 24, 2024, OpenAI also gave the feature a new visual identity, replacing the animated black dots shown at the May demo with a blue animated sphere.[3] Enterprise and Edu customers were to start receiving access the following week.[3] An OpenAI spokesperson said at the time that the feature was not yet available in several regions, including the European Union, the United Kingdom, Switzerland, Iceland, Norway, and Liechtenstein.[3] On September 25, 2024, Altman said the rollout had completed early "except for jurisdictions that require additional external review."[9] In the months before the alpha, OpenAI said it tested GPT-4o's voice capabilities with more than 100 external red teamers speaking 45 languages.[2]

Sky voice controversy

One of the five original ChatGPT voices, named Sky, became the subject of a public dispute in May 2024.[13] After the GPT-4o demo on May 13, 2024, listeners compared Sky to the voice of actress Scarlett Johansson, who had voiced an artificial intelligence assistant in the 2013 film "Her." On the day of the demo, OpenAI chief executive Sam Altman posted the single word "her" on X, which Johansson later described as insinuating that the similarity was intentional.[8][14]

According to a statement Johansson released on May 20, 2024, Altman had approached her the previous September to voice the system, telling her that her involvement could help "bridge the gap between tech companies and creatives." She declined the offer. She said that two days before the May demo, Altman contacted her agent asking her to reconsider, and that the system was released before they could connect. Johansson said she was "shocked, angered and in disbelief that Mr. Altman would pursue a voice that sounded so eerily similar to mine," and that she was "forced to hire legal counsel," whose two letters asked OpenAI to detail how it had created the Sky voice.[8][14]

OpenAI paused the use of Sky in its products as of May 19, 2024.[13] In a statement dated May 20, Altman said, "The voice of Sky is not Scarlett Johansson's, and it was never intended to resemble hers. We cast the voice actor behind Sky's voice before any outreach to Ms. Johansson. Out of respect for Ms. Johansson, we have paused using Sky's voice in our products."[8][13] In a blog post describing how the voices were chosen, published late on May 19 and later updated with a fuller timeline, OpenAI said the voices were selected in a process spanning five months, that a casting call drew more than 400 submissions from voice and screen actors, and that five voices (Breeze, Cove, Ember, Juniper and Sky) were selected.[8][13] It said that "AI voices should not deliberately mimic a celebrity's distinctive voice" and that Sky belonged to a different professional actress using her own natural speaking voice. OpenAI's updated timeline states that Altman spoke with Johansson's team on September 11, 2023, that she declined a week later through her agent, and that he contacted her team again on May 10, 2024.[13]

Later updates (2025)

OpenAI upgraded Advanced Voice for paid users on June 7, 2025, citing subtler intonation, more realistic cadence and more expressiveness for emotions such as empathy and sarcasm, and adding ongoing translation between languages. Its release notes also listed known limitations, including occasional variations in tone and pitch and "rare hallucinations" producing sounds resembling ads, gibberish or background music. The same improvements reached free users on July 17, 2025.[7] On August 7, 2025, OpenAI extended voice to custom GPTs, expanded access to near-unlimited use for Plus users and hours each day for free users, and said Standard Voice Mode would be retired after 30 days; on September 9, 2025, it said it would keep Standard Voice available while it addressed feedback from its users.[7]

The largest interface change came on November 25, 2025, when OpenAI moved ChatGPT Voice into the main chat thread instead of a dedicated full-screen mode. Voice conversations now run inline alongside the text transcript, so users can speak, watch replies appear as text, scroll back through earlier messages, and see visuals such as images, maps, and other on-screen content while talking.[7][10][11] OpenAI's announcement said: "You can talk, watch answers appear, review earlier messages, and see visuals like images or maps in real time."[26] The earlier full-screen experience, an animated blue circle where replies could only be heard rather than read, remained available for users who preferred it through a "Separate mode" toggle in ChatGPT's voice settings, but the in-chat experience became the new default across web and mobile apps.[10][11]

By late 2025, OpenAI's help center also described how voice usage was metered across plans. For subscribers, daily voice use was described as "nearly unlimited," with sessions starting on GPT-4o and switching to GPT-4o mini once the day's GPT-4o minutes were used. Logged-in free users used GPT-4o mini from the start, "subject to a limit of hours each day," and Pro subscribers had unlimited GPT-4o voice subject to abuse guardrails.[12] OpenAI announced on December 11, 2025, that it would retire the voice experience in the ChatGPT macOS app on January 15, 2026, while keeping voice on chatgpt.com, iOS, Android and Windows.[7] Further updates in January and February 2026 improved instruction following and web search in voice.[7]

GPT-Live (2026)

On July 8, 2026, OpenAI introduced GPT-Live, which it called "a new generation of voice models," and began rolling out GPT-Live-1 and GPT-Live-1 mini to ChatGPT users globally on iOS, Android and chatgpt.com.[20] GPT-Live-1 became the default model for ChatGPT Voice for Go, Plus and Pro users and GPT-Live-1 mini the default for free users.[20] OpenAI's release notes said GPT-Live-1 was not available in ChatGPT Business, Enterprise or Edu workspaces at launch and did not support video or screen sharing, which eligible subscribers could keep using through Advanced Voice Mode.[7] At launch, GPT-Live used GPT-5.5 in the background for search and reasoning, and users could choose Instant, Medium or High reasoning levels.[20] OpenAI said that in its human evaluations GPT-Live-1 and GPT-Live-1 mini were "strongly preferred over Advanced Voice Mode," and that more than 150 million people talk to ChatGPT each week using features such as Voice and Dictation.[20] OpenAI later added that, as of July 31, 2026, supported GPT-Live audio carries SynthID watermarking.[20]

Subsequent updates extended the Live experience. ChatGPT Voice became available in Work and Codex in the ChatGPT desktop app on July 23, 2026, where it can start, check and coordinate tasks by voice.[7][25] In August 2026, OpenAI added file uploads and Projects to voice conversations and, on iPhone, Live Activities that show voice content on the Lock Screen and in the Dynamic Island.[7] On September 9, 2026, OpenAI said ChatGPT Voice could use GPT-5.6 or GPT-6 Astra when it needs to search or reason through harder questions, chosen with the same controls as text chat. It also simplified daily limits, stopped switching Plus and Pro users to GPT-Live mini after a limit, and deprecated the separate Instant, Medium and High voice intelligence levels.[7]

September 2026: plugins, GPT-6 models and ChatGPT Work

On September 23, 2026, OpenAI posted that ChatGPT Voice could now "Use plugins like your email, calendar, and Slack," "Be powered by GPT-6 Astra, Sol, and Luna," and be used in ChatGPT Work on web and mobile "so you can create docs, decks, sites, and spreadsheets or tackle complex tasks in the browser, just by talking," rolling out globally that day in the latest version of the app.[23] OpenAI's release notes for the same day describe the plugin feature as the ability to "use the plugins and connected apps available to your account during a Voice conversation" on web, iOS and Android, and say Voice in Work can create documents, presentations and spreadsheets, use connected apps or work in a browser, with unfinished tasks able to continue in text after a call ends.[7] Free and Go users can use Voice in Chat with the plugins their plan supports; Voice in Work requires both Voice and Work access, and existing app connections, permissions and usage limits still apply.[7] According to the help center, actions that need approval must be approved with on-screen controls, because spoken approval is not supported.[22] GPT-6 Sol and GPT-6 Luna had been introduced in ChatGPT Work and Codex the previous day; 9to5Mac described Astra as the most capable of the three and Sol and Luna as medium and small models.[7][24]

Usage limits (as of September 2026)

OpenAI's help center measures Live usage in Chat over a rolling 24-hour period:[22]

PlanVoice in Chat allowance
FreeLimited access to GPT-Live-1 mini; limits may change
Go3 hours with GPT-Live-1 mini
Plus3 hours with GPT-Live-1
Pro ($100/month)15 hours with GPT-Live-1
Pro ($200/month)Unlimited GPT-Live-1
Business Standard3 hours with GPT-Live-1; additional usage at 1.25 credits per minute
Business Premium15 hours with GPT-Live-1; additional usage at 1.25 credits per minute
K-12 Teachers3 hours with GPT-Live-1
Enterprise, Edu, and Clinicians on credit-based pricing1.25 credits per minute
Enterprise on usage-based USD pricing$0.05 per minute
Legacy Enterprise and Edu plans3 hours with GPT-Live-1

Expanded article table

Voice in Work follows the plan's voice limits, and Work tasks started by voice also count toward normal Work usage; voice in the desktop app has separate pricing and limits.[22] Audio clips from Live and Advanced Voice conversations, and video clips from Advanced Voice conversations, are stored with the chat transcript and retained for 30 days.[22]

Reception

Advanced Voice Mode drew substantial attention for the naturalness of its conversations, with TechCrunch describing the voice as "hyperrealistic."[2] The roughly seven-month gap between the May 2024 demo and the December 2024 arrival of live video, a capability shown at the original unveiling, was a recurring point in press coverage.[5] The feature's delayed availability in Europe also generated discussion before the October 2024 EU rollout: Windows Central pointed to commentary that the EU AI Act's restrictions on emotion recognition might explain the delay, while Altman referred only to jurisdictions requiring "additional external review."[9] Johansson used the Sky episode to call for "transparency and the passage of appropriate legislation to help ensure that individual rights are protected."[8] Reviewing the July 2026 GPT-Live demo, TechCrunch reported that the live Hindi translation had a heavy American accent and an unnatural, bookish tone.[21]

References

  1. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9 ^10 ^11 ^12OpenAI, "Hello GPT-4o," May 13, 2024. openai.com/...hello-gpt-4o
  2. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8Maxwell Zeff, "OpenAI releases ChatGPT's hyperrealistic voice to some paying users," TechCrunch, July 30, 2024. techcrunch.com/...ts-super-realistic-voice-feature
  3. ^1 ^2 ^3 ^4 ^5 ^6 ^7Maxwell Zeff, "OpenAI rolls out Advanced Voice Mode with more voices and a new look," TechCrunch, September 24, 2024. techcrunch.com/...-with-more-voices-and-a-new-look
  4. ^1 ^2 ^3Paul Sawers, "OpenAI gives ChatGPT a voice for verbal conversations," TechCrunch, September 25, 2023. techcrunch.com/...openai-chatgpt-voice
  5. ^1 ^2 ^3Kyle Wiggers, "ChatGPT now understands real-time video, seven months after OpenAI first demoed it," TechCrunch, December 12, 2024. techcrunch.com/...ths-after-openai-first-demoed-it
  6. ^1 ^2OpenAI (@OpenAI), post announcing GPT-4o mini-powered Advanced Voice for free users, X, February 25, 2025. x.com/...1894495906952876101
  7. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9 ^10 ^11 ^12 ^13 ^14 ^15 ^16 ^17 ^18 ^19 ^20 ^21 ^22 ^23 ^24 ^25 ^26 ^27 ^28 ^29 ^30 ^31 ^32 ^33OpenAI Help Center, "ChatGPT: Release Notes" (accessed September 25, 2026). help.openai.com/...6825453-chatgpt-release-notes
  8. ^1 ^2 ^3 ^4 ^5 ^6Bobby Allyn, "Scarlett Johansson says she is 'shocked, angered' over new ChatGPT voice," NPR, May 20, 2024. npr.org/...-to-scarlett-johansson-in-the-movie-her
  9. ^1 ^2Kevin Okemwa, "ChatGPT's human-like Advanced Voice Mode feature can say 'Sorry I'm late' in over 50 languages, but the EU is reportedly turning a deaf ear to OpenAI's plea because it can detect user emotion," Windows Central, September 26, 2024. windowscentral.com/...ce-mode-feature-barred-by-eu
  10. ^1 ^2 ^3Sarah Perez, "ChatGPT's voice mode is no longer a separate interface," TechCrunch, November 25, 2025. techcrunch.com/...s-no-longer-a-separate-interface
  11. ^1 ^2 ^3Ian Carlos Campbell, "Now you can use ChatGPT Voice without leaving your chat," Engadget, November 25, 2025. engadget.com/...ithout-leaving-your-chat-195000538
  12. ^1 ^2OpenAI Help Center, "Voice Mode FAQ," archived November 27, 2025 (the original URL now redirects to "ChatGPT Voice"). web.archive.org/...8400625-voice-mode-faq
  13. ^1 ^2 ^3 ^4 ^5 ^6OpenAI, "How the voices for ChatGPT were chosen," May 19, 2024 (updated May 22, 2024). openai.com/...how-the-voices-for-chatgpt-were-chosen
  14. ^1 ^2Aisha Malik, "Scarlett Johansson says that OpenAI approached her to use her voice," TechCrunch, May 20, 2024. techcrunch.com/...-approached-her-to-use-her-voice
  15. ^Pascale Davies, "OpenAI's advanced voice mode is now available in Europe. Here's what you can do with it," Euronews, October 22, 2024. euronews.com/...rope-heres-what-you-can-do-with-it
  16. ^OpenAI, "Navigating the challenges and opportunities of synthetic voices," March 29, 2024. openai.com/...nd-opportunities-of-synthetic-voices
  17. ^1 ^2OpenAI, "Introducing the Realtime API," October 1, 2024. openai.com/...introducing-the-realtime-api
  18. ^OpenAI, "Introducing gpt-realtime and Realtime API updates for production voice agents," August 28, 2025. openai.com/...introducing-gpt-realtime
  19. ^OpenAI, "Changelog," OpenAI API documentation (accessed September 25, 2026). learn.chatgpt.com/...changelog
  20. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9 ^10 ^11 ^12 ^13OpenAI, "Introducing GPT-Live," July 8, 2026 (updated July 31, 2026). openai.com/...introducing-gpt-live
  21. ^Ivan Mehta, "OpenAI releases new voice models for more natural live conversations," TechCrunch, July 8, 2026. techcrunch.com/...-more-natural-live-conversations
  22. ^1 ^2 ^3 ^4 ^5 ^6 ^7OpenAI Help Center, "ChatGPT Voice" (accessed September 25, 2026). help.openai.com/...20001274-chatgpt-voice
  23. ^1 ^2 ^3OpenAI (@OpenAI), post announcing ChatGPT Voice plugins, GPT-6 models and ChatGPT Work support, X, September 23, 2026. x.com/...2102808325742322002
  24. ^Zac Hall, "OpenAI just upgraded ChatGPT Voice in three ways," 9to5Mac, September 23, 2026. 9to5mac.com/...pgraded-chatgpt-voice-in-three-ways
  25. ^ChatGPT Learn, "ChatGPT Voice" (accessed September 25, 2026). learn.chatgpt.com/...voice
  26. ^OpenAI (@OpenAI), post announcing ChatGPT Voice inside chat, X, November 25, 2025. x.com/...1993381101369458763

Improve this article

Add missing citations, update stale details, or suggest a clearer explanation. Every suggestion is reviewed for sourcing before it goes live.

5 revisions · v6 · 4,057 words · full history

Fact-checks are independent of edits: a reviewer re-verifies the article against its sources and stamps the date. How we verify

Research and drafting on this wiki are AI-assisted, under named human editorial standards. How AI is used here

Reviewer note: xg09 I2 audit (14 corrections) + V5 independent verification (~44 sources, ~185 claims across cluster); 3 minor fixes applied

Cite this page: AI Wiki. "Advanced Voice Mode." aiwiki.ai, updated 25 Sept 2026, fact-checked 25 Sept 2026. CC BY 4.0. https://aiwiki.ai/wiki/advanced_voice_mode

Suggest edit