SendTech Times
العربية
AINews|July 9, 2026, 03:09 PM
AI SHIFT:

OpenAI Rolls Out GPT-Live For ChatGPT Voice Without API Timing

BySendTech Times AI & Enterprise DeskNewsroom-edited, source-reviewed coverage|Source: OpenAI
Newsroom brief

OpenAI said GPT-Live will let ChatGPT Voice listen and speak at the same time, with GPT-5.5 handling harder search and reasoning in the background. The company cited more than 150 million weekly Voice and Dictation users but did not give API timing, video support dates or independent benchmark validation.

Verified against source materialEdited by SendTech Times AI & Enterprise Desk
OpenAI Rolls Out GPT-Live For ChatGPT Voice Without API Timing
Image source: OpenAI

OpenAI is moving ChatGPT Voice to GPT-Live, a full-duplex voice model that the company says can listen and speak at the same time, while delegating harder questions to GPT-5.5 in the background.

The GPT-Live ChatGPT Voice rollout starts globally for ChatGPT users, but OpenAI did not include API launch timing, video support or screen-sharing availability at launch.

The source is OpenAI's own company announcement, so performance, safety and rollout claims remain company-provided and not independently verified.

GPT-Live Rolls Out To ChatGPT Voice Users Globally

OpenAI said it is beginning to roll out two versions of the new system, GPT-Live-1 and GPT-Live-1 mini, to ChatGPT users globally.

The company said GPT-Live-1 will become the default model for ChatGPT Voice on Go, Plus and Pro plans, while GPT-Live-1 mini will become the default for Free users.

The rollout covers ChatGPT on iOS, Android and ChatGPT.com.

OpenAI said it plans to bring GPT-Live to the API soon, but the announcement directed developers and enterprises to a notification form rather than a release date.

The company framed GPT-Live as a change in voice architecture rather than only a new interface.

Earlier cascaded systems chained speech-to-text, language-model response and text-to-speech models.

OpenAI said that approach could lose information across models and produce slow, stilted responses.

Full-Duplex Voice Lets The Model Listen And Speak Together

OpenAI said GPT-Live uses a full-duplex architecture, meaning the model continuously processes input while generating output.

The company said the model can decide many times per second whether to speak, keep listening, pause, interrupt or use a tool.

That design is meant to reduce one of the common limits in voice assistants: rigid turn-taking.

OpenAI said turn-based voice models wait for a user to stop speaking before responding, and silence-based turn detection can mistake a short pause or background noise for the end of a turn.

In the new ChatGPT Voice experience, OpenAI said users can interrupt with a question, pause while thinking or ask ChatGPT to stay quiet and listen.

The company also said the model can acknowledge that it is listening with short phrases such as "mhmm" or "got it", and that it has remastered the nine distinct voices in ChatGPT for GPT-Live.

GPT-5.5 Handles Search And Reasoning Behind Voice Replies

OpenAI said GPT-Live keeps the live audio exchange separate from heavier tasks.

For questions involving search, reasoning or more agent-like work, the company said GPT-Live can hand the request to another model such as GPT-5.5 while the spoken session continues.

At launch, OpenAI said GPT-Live will use GPT-5.5 in the background.

The company also said GPT-Live-1 Instant and GPT-Live-1 mini use GPT-5.5 Instant, while GPT-Live-1 Medium and GPT-Live-1 High use GPT-5.5 Thinking with medium and high reasoning effort.

The company said the voice mode can now show rich visual cards during spoken conversations for topics such as weather, stocks and sports.

It also said Voice continues to support search, memory, images and file uploads.

OpenAI Cites 150 Million Weekly Voice And Dictation Users

OpenAI said more than 150 million people each week talk to ChatGPT using features such as Voice and Dictation.

The company said GPT-Live-1 and GPT-Live-1 mini were strongly preferred over Advanced Voice Mode in matched 5-10 minute human evaluations covering preference, turn-taking, interruptions, conversational flow and naturalness.

Those evaluation claims remain company-provided.

OpenAI said GPT-Live-1 outperforms Advanced Voice Mode on GPQA, BrowseComp and an internal tau3-Voice Telecom variant, but the announcement did not include independent benchmark validation or customer-level enterprise deployment data.

The safety section also stays inside OpenAI's own testing boundary.

The company said GPT-Live adds audio-native evaluations, synthetic audio testing, internal red-teaming, safeguards for higher-risk conversations, teen protections and voice-impersonation limits based on predefined voices.

OpenAI did not disclose API release timing, pricing for developer access, independent benchmark results, enterprise customer deployments, support dates for voice with video or screen sharing, or language-by-language quality data for GPT-Live.

Share this article
inXf

Related articles

More
Google Adds Gemini Agent To Search Ads In India Beta
AI

Google Adds Gemini Agent To Search Ads In India Beta

Google launched Business Agent for Leads in India as a Gemini-powered Search ad format that can chat with users on the search page. The company cited $77.25 billion in first-quarter ad revenue and several India ad metrics, but did not disclose pricing, wider rollout dates or independent lead-quality validation.

Grep Adds LLM Agent To Monito As Online Proctoring Shifts Toward Context Review
AI

Grep Adds LLM Agent To Monito As Online Proctoring Shifts Toward Context Review

Grep said its Monito online proctoring product now uses an LLM agent to analyze context around suspected cheating events. The company cited internal tests showing more than 30 percent shorter post-exam review time and nearly 20 percent fewer false alerts. The key issue is whether agent-based proctoring can improve review efficiency while preserving human final judgment and candidate fairness.

HP Moves OpenAI Frontier From Pilots To Enterprise Workflows Without ROI Data
AI

HP Moves OpenAI Frontier From Pilots To Enterprise Workflows Without ROI Data

HP is scaling an OpenAI Frontier partnership across support, security, device and software workflows after pilots, but the companies did not disclose pricing, deployment dates or measured return on investment.

OpenSearch Foundation Pitches AI Data Layer Without Customer ROI Proof
AI

OpenSearch Foundation Pitches AI Data Layer Without Customer ROI Proof

Linux Foundation executives described OpenSearch as an open data layer for AI search, observability and security monitoring as agents push query scale higher. The interview cited 100,000 queries within a minute and security scans across 152 repositories, but did not name customer deployments or audited savings.

NVIDIA Agent Toolkit Adds Runtime Controls But No Rollout Counts
AI

NVIDIA Agent Toolkit Adds Runtime Controls But No Rollout Counts

NVIDIA is packaging Nemotron open models, NemoClaw blueprints and OpenShell runtime support for specialized enterprise agents, but the company did not disclose pricing, deployment dates or rollout counts.

Fujitsu pairs OpenAI and Anthropic deals for enterprise AI push
AI

Fujitsu pairs OpenAI and Anthropic deals for enterprise AI push

Fujitsu announced cooperation with OpenAI and a strategic partnership with Anthropic on the same day. The company plans to use ChatGPT Enterprise, Codex and Claude internally and in customer-facing enterprise AI services. Fujitsu says it will combine external models with Fujitsu Kozuchi, Takane, cybersecurity work and a Forward Deployed Engineer model.

Keep Reading

More Stories

Latest
Malaysia Chip Outlook Rides AI Demand But Advanced Packaging Gap RemainsChips & SemiconductorsJul 9, 2026Malaysia Chip Outlook Rides AI Demand But Advanced Packaging Gap RemainsMaybank, Kenanga, MBSB and TA Securities said AI and data centre demand is supporting Malaysia’s technology sector, while Kenanga said exposure to higher-value advanced packaging remains limited. TA cited MYR 59.9 billion in semiconductor-related investments under the National Semiconductor Strategy.Bitcoin Suisse Wins ADGM Licence For Institutional Crypto ServicesCapital & PolicyJul 9, 2026Bitcoin Suisse Wins ADGM Licence For Institutional Crypto ServicesBitcoin Suisse has received an ADGM Financial Services Permission for its Abu Dhabi subsidiary BTCS (Middle East) Ltd., allowing custody, approved virtual-asset trading and portfolio management for institutional and professional investors. The firm cited $3.7 billion in crypto assets under custody, but the announcement did not name UAE clients, launch volumes or retail availability.Planno Takes Solar GeoAI To 16 Markets Without Investment AmountAIJul 9, 2026Planno Takes Solar GeoAI To 16 Markets Without Investment AmountDubai-based Planno says Incubayt backing helped expand its solar GeoAI platform into 16 markets, while its Dubai rooftop analysis lists 4 GW of unbuilt capacity and no disclosed investment amount.Open USD Stablecoin Draws 140 Backers But Launch Timing Stays OpenFintech & Digital PaymentsJul 9, 2026Open USD Stablecoin Draws 140 Backers But Launch Timing Stays OpenOpen Standard launched Open USD with backing from banks, card networks and fintech firms, while the consortium said the stablecoin will go live later this year without naming commercial volumes or merchant adoption targets.National Grid Backs Joulent With $1.75 Billion As AI Power Queues StretchCloud & Data CentersJul 9, 2026National Grid Backs Joulent With $1.75 Billion As AI Power Queues StretchNational Grid Ventures plans a $1.75 billion minority investment in Joulent, a power developer built around dedicated generation for AI data centres. The deal supports Project Kilby in Texas, but delivery and interconnection risks remain outside the disclosed timetable.Bhashini Opens Offline AI Hackathon For Indian Government PilotsAIJul 9, 2026Bhashini Opens Offline AI Hackathon For Indian Government PilotsBhashini, Current AI and Kalpa Impact are inviting teams to build offline, multilingual AI devices for Indian public services, but the programme still has to prove demand, funding and long-term maintenance.Fed Proposes Risk-Based AML Rule For Banks With 60-Day Comment WindowCapital & PolicyJul 9, 2026Fed Proposes Risk-Based AML Rule For Banks With 60-Day Comment WindowThe Federal Reserve Board said its proposed risk-based AML/CFT requirements would apply to supervised banks, with comments due 60 days after Federal Register publication and no final implementation date yet.Texas Approves $13.8 Billion 765 kV Grid Plan For AI Data Centre LoadsCloud & Data CentersJul 9, 2026Texas Approves $13.8 Billion 765 kV Grid Plan For AI Data Centre LoadsTexas regulators approved a 765 kV transmission plan estimated by ERCOT at $13.8 billion, but the plan does not name AI data centre customers, final routes or ratepayer cost terms.PitchBook Says AI Took $355.9 Billion Of US Venture Deals In First HalfCapital & PolicyJul 9, 2026PitchBook Says AI Took $355.9 Billion Of US Venture Deals In First HalfThe PitchBook-NVCA Venture Monitor put US venture deal value at $412.7 billion in the first half of 2026 and said AI companies took $355.9 billion of that total. The report also warned that returns could narrow if AI growth or returns disappoint.Imec Says AI Inference Pushes Optical I/O Closer To ChipsChips & SemiconductorsJul 9, 2026Imec Says AI Inference Pushes Optical I/O Closer To ChipsImec researchers told EE Times that AI inference is shifting the interconnect problem from rack-scale optics towards 2.5D and 3D optical I/O near processors. The group discussed a 250 Tb/s bandwidth projection but did not disclose commercial product timing, customer deployments or cooling test results.UAE AI Infrastructure Draws 29 Percent Of Greenfield FDI Capital SpendingCapital & PolicyJul 9, 2026UAE AI Infrastructure Draws 29 Percent Of Greenfield FDI Capital SpendingUAE Investment Minister Mohamed Hassan Alsuwaidi said communications projects including data centres and AI infrastructure accounted for 29 percent of greenfield FDI capital expenditure in 2025. The ministry report also put inward FDI at $48.3 billion, while named cloud operators, project-level financing and delivery dates were not disclosed.Samsung Starts PM1763 PCIe 6.0 SSD Production For AI ServersChips & SemiconductorsJul 9, 2026Samsung Starts PM1763 PCIe 6.0 SSD Production For AI ServersSamsung has started mass production of the PM1763, a PCIe 6.0 enterprise SSD for AI and HPC servers with 28,400 MB/s sequential reads in a 16TB configuration, according to ServeTheHome. The report did not name customers, pricing, shipment volumes or server qualification dates.