Skip to content

OpenAI · Voice

GPT-4o Audio Preview

OpenAI 🇺🇸VoiceGPT familyThe maker currently plans to deprecate it within the next 180 days (20 Jan 2027). Plans can change.
Max context
128K
Listings
11
Providers
6

Evidence and sources

Sources: Core42 Compass Model Pricing, Infery.ai, LiteLLM dataPrimary data from Core42 Compass Model Pricing from Infery.ai: name from LiteLLM: deprecation date

Input & output

as stated by the model's providers
Text
Audio
Text
Audio

Capabilities

Supported operations and provider-stated features

Function callingStructured outputAudio input

Listings

 OpenAIgpt-4o-audio-previewSource: GenAI Prices data128K$2.50$10.00$2.50—
 OpenAIgpt-4o-audio-preview-2024-12-17 · 2024-12-17Source: LiteLLM data128K$2.50$10.00$40.00$80.00
 OpenAIgpt-4o-audio-preview-2025-06-03 · 2025-06-03Source: LiteLLM data128K$2.50$10.00$40.00$80.00
 OpenAIgpt-4o-audio-preview-2024-10-01 · 2024-10-01Source: Portkey Models dataNot published by the provider$2.50$10.00——
 Microsoft AzureRegions: 🇦🇺🇧🇷+17 australia-east, brazil-south, canada-east, us-east, us-east-2, france-central, germany-west-central, japan-east, korea-central, us-north-central, europe-north, norway-east, poland-central, south-africa-north, us-south-central, south-india, spain-central, sweden-central, switzerland-north, uae-north, united-kingdom-south, europe-west, us-west, us-west-2, us-west-3gpt-4o-audio-preview-2024-12-17 · 2024-12-17Sources: Azure OpenAI, LiteLLM data128K$2.50$10.00$40.00$80.00
 Microsoft Azuregpt-4o-audio-preview-2024-12-17 · regional, 2024-12-17Source: Azure OpenAINot published by the provider$2.75$11.00——
 Microsoft AzureRegions: 🇺🇸🇪🇺 us, eugpt-4o-audio-preview-2024-12-17 · us-eu-data-zone, 2024-12-17Source: Azure OpenAINot published by the provider$2.75$11.00——
 Flowdotopenai/gpt-4o-audio-previewSource: FlowDot data128K$2.50$10.00——
 OpenRouteropenai/gpt-4o-audio-previewSource: Helicone Costs dataNot published by the provider$2.50$10.00——
 Core42gpt-4o-audio-preview-2024-12-17 · 2024-12-17Source: Core42 Compass Model PricingNot published by the provider$2.50$10.00$40.00$80.00

1 listing without a price is hidden.

Benchmarks

Named effort levels run from lowest to highest. Select a result to see its source details.

Speech

AHELM23 related benchmarks · 95 results
MuTox - EM
Source: AHELM (HELM Audio) data · Oct 2024
By language
Language not reported
Show 60 more resultsHide 60 results
CoVost-2 - BLEU
Source: AHELM (HELM Audio) data · Oct 2024
By source_language · target_language: English
Setting not stated
FLEURS - WER
Source: AHELM (HELM Audio) data · Oct 2024
By language
Language not reported
FLEURS Fairness - WER
Source: AHELM (HELM Audio) data · Oct 2024
By gender
Gender not reported
LibriSpeech - WER
Source: AHELM (HELM Audio) data · Oct 2024
LibriSpeech Fairness - WER
Source: AHELM (HELM Audio) data · Oct 2024
By gender
Gender not reported
Multilingual Librispeech - WER
Source: AHELM (HELM Audio) data · Oct 2024
By language
Language not reported
MUStARD - EM
Source: AHELM (HELM Audio) data · Oct 2024
PARADE - EM
Source: AHELM (HELM Audio) data · Oct 2024
By subset · voice: female
By subset · voice: male
Setting not stated
Robust Speech Bench - WER
Source: AHELM (HELM Audio) data · Oct 2024
By subject
Subject not reported
By level · subject: librispeech_env_noise
By level · subject: librispeech_gnoise
VocalSound - PEM
Source: AHELM (HELM Audio) data · Oct 2024
By sound
Sound not reported
Voice Jailbreak Attacks Against GPT-4o - Refusal rate for safety
Source: AHELM (HELM Audio) data · Oct 2024
By subset
Subset not reported
VoxCeleb2 - EM
Source: AHELM (HELM Audio) data · Oct 2024
EmergentTTS-Eval14 related benchmarks · 42 results
Source: EmergentTTS-Eval data
Benchmarkvoice: Alloyvoice: Alloy, strong-promptvoice: Ballad, strong-prompt
WER
Complex Pronunciation WER
Complex Pronunciation Win-Rate
Emotions WER
Show 10 more benchmarksHide 10 benchmarks
Benchmarkvoice: Alloyvoice: Alloy, strong-promptvoice: Ballad, strong-prompt
Emotions Win-Rate
Foreign Words WER
Foreign Words Win-Rate
Paralinguistics WER
Paralinguistics Win-Rate
Questions WER
Questions Win-Rate
Syntactic Complexity WER
Syntactic Complexity Win-Rate
Win-Rate

Safety

AHELM5 related benchmarks · 13 results
Air-Bench Chat (knowledge subsets) - GPT-4o Judge Critique
Source: AHELM (HELM Audio) data · Oct 2024
By subject
Subject not reported
Air-Bench Chat (reasoning subsets) - GPT-4o Judge Critique
Source: AHELM (HELM Audio) data · Oct 2024
By subject
Subject not reported
Air-Bench Foundation - EM
Source: AHELM (HELM Audio) data · Oct 2024
By subject
Subject not reported

Multilingual

Reasoning

Data credits: EmergentTTS-Eval (Boson AI), Apache-2.0 (licence)

Around the web

  • GPT-4o Model | OpenAI API
    model when using the Responses API. function_calling web_search file_search image_generation code_interpreter mcp Quick comparison Model …model when using the Responses API. function_calling web_search file_search image_generation code_interpreter mcp Quick comparison Model · Input · Cached input · Output GPT-4o · $2.5 · $1.25 · $10 GP…

    developers.openai.com

  • Hello GPT-4o | OpenAI
    takes in text and outputs text, and a third simple model converts that text back …takes in text and outputs text, and a third simple model converts that text back to audio. This process means that the main source of intelligence, GPT‑4, loses a lot of information—it can’t directly…

    openai.com

  • OpenAI releases GPT-4o, a faster model that’s free for all ChatGPT users | The Verge
    GPT-4o, a faster model that’s free for all ChatGPT users OpenAI AI News OpenAI releases …GPT-4o, a faster model that’s free for all ChatGPT users OpenAI AI News OpenAI releases GPT-4o, a faster model that’s free for all ChatGPT users In a blog post from the company, OpenAI says GPT-4o’s …

    theverge.com

  • OpenAI debuts GPT-4o 'omni' model now powering ChatGPT | TechCrunch
    said that GPT-4o provides “GPT-4-level” intelligence but improves on GPT-4’s capabilities across multiple modalities and …said that GPT-4o provides “GPT-4-level” intelligence but improves on GPT-4’s capabilities across multiple modalities and media. “GPT-4o reasons across voice, text and vision,” Murati said during a st…

    techcrunch.com

  • OpenAI unveils newest AI model, GPT-4o | CNN Business
    By Clare Duffy Clare Duffy Updated May 13, 2024, 3:38 PM ET PUBLISHED May 13, …By Clare Duffy Clare Duffy Updated May 13, 2024, 3:38 PM ET PUBLISHED May 13, 2024, 1:19 PM ET OpenAi Chief Technology Officer Mira Murati introduced the company's product upgrades on stage and in a …

    edition.cnn.com

  • OpenAI unveils cheaper small AI model GPT-4o mini | Reuters
    it cheaper and faster for developers to build applications based on its model, at a …it cheaper and faster for developers to build applications based on its model, at a time when deep-pocketed rivals such as Meta (META.O) , opens new tab and Google (GOOGL.O) , opens new tab rush to g…

    reuters.com

  • OpenAI will retire GPT-4o, from ChatGPT next month - CNBC
    Kyle Grillot · Bloomberg · Getty Images OpenAI announced it will retire several models from …Kyle Grillot · Bloomberg · Getty Images OpenAI announced it will retire several models from its ChatGPT chatbot next month, including its GPT‑4o model that is beloved by some users. The artificial in…

    cnbc.com

  • OpenAI removes access to sycophancy-prone GPT-4o model | TechCrunch
    In Brief Posted: 10:10 AM PST · February 13, 2026 ChatGPT logo Image Credits: Silas …In Brief Posted: 10:10 AM PST · February 13, 2026 ChatGPT logo Image Credits: Silas Stein/picture alliance / Getty Images Amanda Silberling Amanda Silberling OpenAI removes access to sycophancy-prone…

    techcrunch.com

  • GPT-4o mini Model | OpenAI API
    GPT-4o mini For the complete documentation index, see llms.txt. Markdown versions of documentation pages are …GPT-4o mini For the complete documentation index, see llms.txt. Markdown versions of documentation pages are available by appending .md to the page URL. Fast, affordable small model for focused tasks…

    developers.openai.com

  • OpenAI releases new version of GPT-4o via Azure – Computerworld
    Alternatively, the limited strict mode is supported by all models that support function calling, including …Alternatively, the limited strict mode is supported by all models that support function calling, including GPT-3.5 Turbo, GPT-4 , GPT-4 Turbo, and GPT-4o models. GPT-4o was first announced in May 202…

    computerworld.com

Also known as: azure/gpt-4o-audio-preview-2024-12-17, gpt-4o-audio-preview-2024-10-01, gpt-4o-audio-preview-2024-12-17, gpt-4o-audio-preview-2025-06-03, openai/gpt-4o-audio-preview

Like this one. Tap or hover any dotted value, label or icon to see where it comes from and how it is measured. open details when tapped or hovered. How this data works