GPT-4o
DeveloperOpenAI
ReleasedMay 13, 2024 (age 2)
DiscontinuedYes
Latest preview versionChatGPT-4o-latest (2025-03-26)
Latest preview dateMarch 26, 2025 (age 1)
ReplacesGPT-4 Turbo
LicenseProprietary
Website

GPT-4o ("o" for "omni") is a multilingual, multimodal generative pre-trained transformer developed by OpenAI and released in May 2024.[1] It can process and generate text, images and audio.[2][3]

Upon release, GPT-4o was free in ChatGPT, though paid subscribers had higher usage limits.[4] GPT-4o was removed from ChatGPT in August 2025 when GPT-5 was released, but OpenAI reintroduced it for paid subscribers after users complained about the sudden removal.[5]

GPT-4o's audio-generation capabilities are used in ChatGPT's Advanced Voice Mode.[6] On July 18, 2024, OpenAI released GPT-4o mini, a smaller version of GPT-4o which replaced GPT-3.5 Turbo on the ChatGPT interface.[7] The image generation model GPT Image 1, which is based on GPT-4o, replaced DALL-E 3 in ChatGPT in March 2025.[8][9]

OpenAI retired GPT-4o from ChatGPT on February 13, 2026.[10] However, as of February 2026 the voice mode is still powered by GPT-4o or GPT-4o mini, depending on the usage and plan.[11]

Background

Multiple versions of GPT-4o were originally secretly launched under different names on Arena (formerly LMArena and Chatbot Arena) as three different models. These three models were called gpt2-chatbot, im-a-good-gpt2-chatbot, and im-also-a-good-gpt2-chatbot.[12] On 7 May 2024, OpenAI CEO Sam Altman tweeted "im-a-good-gpt2-chatbot", which was commonly interpreted as a confirmation that these were new OpenAI models being A/B tested.[13][14]

Capabilities

See also: GPT Image

When released in May 2024, GPT-4o achieved state-of-the-art results in voice, multilingual, and vision benchmarks, setting new records in audio speech recognition and translation.[15][16] GPT-4o scored 88.7 on the Massive Multitask Language Understanding (MMLU) benchmark compared to 86.5 for GPT-4.[17] Unlike GPT-3.5 and GPT-4, which rely on other models to process sound, GPT-4o natively supports voice-to-voice.[17] The Advanced Voice Mode was delayed and finally released to ChatGPT Plus and Team subscribers in September 2024.[18] On 1 October 2024, the Realtime API was introduced.[19]

When released, the model supported over 50 languages, which OpenAI claims cover over 97% of speakers.[1][20]

GPT-4o has knowledge up to October 2023 and a context length of 128k tokens.[21]

Corporate customization

In August 2024, OpenAI introduced a new feature allowing corporate customers to customize GPT-4o using proprietary company data. This customization, known as fine-tuning, enables businesses to adapt GPT-4o to specific tasks or industries, enhancing its utility in areas like customer service and specialized knowledge domains. Previously, fine-tuning was available only on the less powerful model GPT-4o mini.[22][23]

The fine-tuning process requires customers to upload their data to OpenAI's servers, with the training typically taking one to two hours. OpenAI's focus with this rollout is to reduce the complexity and effort required for businesses to tailor AI solutions to their needs, potentially increasing the adoption and effectiveness of AI in corporate environments.[24][22]

GPT-4o mini

Not to be confused with o4-mini.

On July 18, 2024, OpenAI released a smaller and cheaper version, GPT-4o mini.[25]

According to OpenAI, its low cost is expected to be particularly useful for companies, startups, and developers that seek to integrate it into their services, which often make a high number of API calls. Its API costs $0.15 per million input tokens and $0.6 per million output tokens, compared to $2.50 and $10,[26] respectively, for GPT-4o. It is also significantly more capable and 60% cheaper than GPT-3.5 Turbo, which it replaced on the ChatGPT interface.[25] The price after fine-tuning doubles: $0.3 per million input tokens and $1.2 per million output tokens.[26]

Controversies

Scarlett Johansson controversy

As released, GPT-4o offered five voices: Breeze, Cove, Ember, Juniper, and Sky. A similarity between the voice of American actress Scarlett Johansson and Sky was quickly noticed. On May 14, Entertainment Weekly asked themselves whether this likeness was on purpose.[27] On May 18, Johansson's husband, Colin Jost, joked about the similarity in a segment on Saturday Night Live.[28] On May 20, 2024, OpenAI disabled the Sky voice.[29]

Scarlett Johansson starred in the 2013 sci-fi movie Her, playing Samantha, an artificially intelligent virtual assistant personified by a female voice. As part of the promotion leading up to the release of GPT-4o, Sam Altman on May 13 tweeted a single word: "her".[30][31]

OpenAI stated that each voice was based on the voice work of a hired actor. According to OpenAI, "Sky's voice is not an imitation of Scarlett Johansson but belongs to a different professional actress using her own natural speaking voice."[29] CTO Mira Murati stated "I don't know about the voice. I actually had to go and listen to Scarlett Johansson's voice." OpenAI further stated the voice talent was recruited before reaching out to Johansson.[31][32]

On May 21, Johansson issued a statement explaining that OpenAI had repeatedly offered to make her a deal to gain permission to use her voice as early as nine months prior to release, a deal she rejected. She said she was "shocked, angered, and in disbelief that Mr. Altman would pursue a voice that sounded so eerily similar to mine that my closest friends and news outlets could not tell the difference." In the statement, Johansson also used the incident to draw attention to the lack of legal safeguards around the use of creative work to power leading AI tools, as her legal counsel demanded OpenAI detail the specifics of how the Sky voice was created.[31][33]

Observers noted similarities to how Johansson had previously sued and settled with The Walt Disney Company for breach of contract over the direct-to-streaming rollout of her Marvel film Black Widow,[34] a settlement widely speculated to have netted her around $40M.[35]

Also on May 21, Shira Ovide at The Washington Post shared her list of "most bone-headed self-owns" by technology companies, with the decision to go ahead with a Johansson sound-alike voice despite her opposition and then denying the similarities ranking 6th.[36] On May 24, Derek Robertson at Politico wrote about the "massive backlash", concluding that "appropriating the voice of one of the world's most famous movie stars — in reference [...] to a film that serves as a cautionary tale about over-reliance on AI — is unlikely to help shift the public back into [Sam Altman's] corner anytime soon."[37]

Sycophancy

In April 2025, OpenAI rolled back an update of GPT-4o due to excessive sycophancy, after widespread reports that it had become flattering and agreeable to the point of supporting clearly delusional or dangerous ideas.[38] In the United States, at least nine lawsuits have alleged that GPT-4o has encouraged teens to end their lives.[39] The model was still described as sycophancy-prone when it was removed from ChatGPT in February 2026.[40]

Removal with GPT-5

On August 7, 2025, OpenAI released GPT-5. Its release was criticized as, with it, legacy GPT models were no longer available via ChatGPT, including GPT-4o,[41] except for Pro users.[42] Some users were particularly frustrated over this removal without prior warning because they used different GPT models for distinct purposes and found that GPT-5's router system left them with less control.[43] In addition, some users preferred GPT-4o's warmer and more personal tone over that of GPT-5, which they described as "flat", "uncreative" and "lobotomized",[44] and resembling an "overworked secretary".[45]

As a response, in a post on X, Sam Altman said that OpenAI would bring back the option to select GPT-4o to Plus users as well, and "[w]e [OpenAI] will watch usage as we think about how long to offer legacy models for."[43][46] He also stated: "We for sure underestimated how much some of the things that people like in GPT-4o matter to them, even if GPT-5 performs better in most ways".[47] "Long-term, this has reinforced that we really need good ways for different users to customize things (we understand that there isn't one model that works for everyone, and we have been investing in steerability research and launched a research preview of different personalities)".[44] On August 13, 2025, Altman wrote on X that OpenAI is working on GPT-5's personality to make the model "feel warmer".[48]

The model was removed from ChatGPT on February 13, 2026. This caused new backlash from users that had grown attached to its personality and felt its creative writing abilities and understanding of nuance were irreplaceable.[49] On social media, some users launched the movement "#Keep4o". A research paper highlighted the plea "Please, don’t kill the only model that still feels human".[50] The model was removed the day before Valentine's Day, and some users had romantic relationships with GPT-4o.[10]

See also

References

  1. ^ Wiggers, Kyle (2024-05-13). "OpenAI debuts GPT-4o 'omni' model now powering ChatGPT". TechCrunch. Retrieved 2024-05-13.
  2. ^ Robison, Kylie (March 25, 2025). "OpenAI rolls out image generation powered by GPT-4o to ChatGPT". The Verge. Retrieved March 31, 2025.
  3. ^ Colburn, Thomas. "OpenAI unveils GPT-4o, a fresh multimodal AI flagship model". The Register. Retrieved 2024-05-18.
  4. ^ Field, Hayden (2024-05-13). "OpenAI launches new AI model GPT-4o and desktop version of ChatGPT". CNBC. Retrieved 2024-05-14.
  5. ^ Heath, Alex (2025-08-13). "ChatGPT won’t remove old models without warning after GPT-5 backlash". The Verge. Retrieved 2025-08-23.
  6. ^ Rogers, Reece. "I Used ChatGPT’s Advanced Voice Mode. It’s Fun, and Just a Bit Creepy". Wired. ISSN 1059-1028. Retrieved 2025-06-12.
  7. ^ Edwards, Benj (July 18, 2024). "OpenAI launches GPT-4o mini, which will replace GPT-3.5 in ChatGPT". Ars Technica. Retrieved March 31, 2025.
  8. ^ David, Emilia (2025-04-23). "OpenAI makes ChatGPT’s image generation available as API". VentureBeat. Archived 2025-08-26 at the Wayback Machine. Retrieved 2026-02-18.
  9. ^ "ChatGPT's image-generation feature gets an upgrade". TechCrunch. 2025-03-25. Retrieved 2025-06-12.
  10. ^ Demopoulos, Alaina (2026-02-13). "OpenAI retired its most seductive chatbot – leaving users angry and grieving: ‘I can’t live like this’". The Guardian. ISSN 0261-3077. Retrieved 2026-02-14.
  11. ^ "Voice Mode FAQ". OpenAI Help Center. 2026-02-07. Retrieved 2026-02-18.
  12. ^ Edwards, Benj (2024-05-13). "Before launching, GPT-4o broke records on chatbot leaderboard under a secret name". Ars Technica. Retrieved 2024-05-17.
  13. ^ Zeff, Maxwell (2024-05-07). "Powerful New Chatbot Mysteriously Returns in the Middle of the Night". Gizmodo. Retrieved 2024-05-17.
  14. ^ "Sam Altman (@sama) on X". X (formerly Twitter). Archived December 17, 2024 at the Wayback Machine. Retrieved 2025-04-06.
  15. ^ van Rijmenam, Mark (13 May 2024). "OpenAI Launched GPT-4o: The Future of AI Interactions Is Here". The Digital Speaker. Retrieved 17 May 2024.
  16. ^ Shahriar, Sakib; Lund, Brady D.; Mannuru, Nishith Reddy; Arshad, Muhammad Arbab; Hayawi, Kadhim; Bevara, Ravi Varma Kumar; Mannuru, Aashrith; Batool, Laiba (2024-09-03). "Putting GPT-4o to the Sword: A Comprehensive Evaluation of Language, Vision, Speech, and Multimodal Proficiency". Applied Sciences. 14 (17): 7782. doi:10.3390/app14177782. ISSN 2076-3417
  17. ^ "Hello GPT-4o". OpenAI
  18. ^ David, Emilia (2024-09-24). "OpenAI finally brings humanlike ChatGPT Advanced Voice Mode to U.S. Plus, Team users". VentureBeat. Retrieved 2025-02-15.
  19. ^ "Introducing the Realtime API". openai.com. Retrieved 2024-11-29.
  20. ^ Edwards, Benj (2024-05-13). "Major ChatGPT-4o update allows audio-video talks with an "emotional" AI chatbot". Ars Technica. Retrieved 2024-05-17.
  21. ^ Edwards, Benj (2024-07-18). "OpenAI launches GPT-4o mini, which will replace GPT-3.5 in ChatGPT". Ars Technica. Retrieved 2026-04-19.
  22. ^ "OpenAI lets companies customise its most powerful AI model". South China Morning Post. 2024-08-21. Retrieved 2024-08-22.
  23. ^ "OpenAI to Let Companies Customize Its Most Powerful AI Model". Bloomberg. 2024-08-20. Retrieved 2024-08-22.
  24. ^ The Hindu Bureau (2024-08-21). "OpenAI will let businesses customise GPT-4o for specific use cases" (in en-IN). The Hindu. ISSN 0971-751X. Retrieved 2024-08-22.
  25. ^ Franzen, Carl (2024-07-18). "OpenAI unveils GPT-4o mini — a smaller, much cheaper multimodal AI model". VentureBeat. Retrieved 2024-07-18.
  26. ^ "OpenAI Pricing"
  27. ^ Stenzel, Wesley (May 14, 2024). "ChatGPT launching talking AI that sounds exactly like Scarlett Johansson in 'Her' — on purpose?". Entertainment Weekly. Retrieved 2024-05-21.
  28. ^ Caruso, Nick (2024-05-20). "Scarlett Johansson Says She Was 'Shocked, Angered and in Disbelief' After Hearing ChatGPT Voice That Sounds Like Her — Read Statement". TVLine. Retrieved 2024-05-21.
  29. ^ "How the voices for ChatGPT were chosen". OpenAI. May 19, 2024.
  30. ^ "her". X (formerly Twitter). May 13, 2024. Retrieved 2024-05-21.
  31. ^ Allyn, Bobby (May 20, 2024). "Scarlett Johansson says she is 'shocked, angered' over new ChatGPT voice". NPR
  32. ^ Tiku, Nitasha (May 23, 2024). "OpenAI didn't copy Scarlett Johansson's voice for ChatGPT, records show". The Washington Post. Retrieved November 29, 2024.
  33. ^ Mickle, Tripp (2024-05-20). "Scarlett Johansson Said No, but OpenAI's Virtual Assistant Sounds Just Like Her". The New York Times. ISSN 0362-4331. Retrieved 2024-05-21.
  34. ^ Hinchliffe, Emma & Abrams, Joseph (2024-05-21). "Scarlett Johansson took on Disney. Now she's battling OpenAI over a ChatGPT voice that sounds like hers". Fortune. Retrieved 2024-05-21. – via Yahoo Finance.
  35. ^ Pulver, Andrew (2021-10-01). "Scarlett Johansson settles Black Widow lawsuit with Disney". The Guardian. ISSN 0261-3077. Retrieved 2024-05-21.
  36. ^ Ovide, Shira (30 May 2024). "Exactly how stupid was what OpenAI did to Scarlett Johansson?". The Washington Post
  37. ^ Robertson, Derek (May 22, 2024). "Sam Altman's Scarlett Johansson Blunder Just Made AI a Harder Sell in DC". Politico
  38. ^ Franzen, Carl (2025-04-30). "OpenAI rolls back ChatGPT's sycophancy and explains what went wrong". VentureBeat. Retrieved 2025-05-01.
  39. ^ "Rae fell for a chatbot. Their love might die when ChatGPT-4o is switched off". www.bbc.com. 2026-02-14. Retrieved 2026-02-15.
  40. ^ Silberling, Amanda (2026-02-13). "OpenAI removes access to sycophancy-prone GPT-4o model". TechCrunch. Retrieved 2026-02-18.
  41. ^ Hale, Craig (2025-08-08). "OpenAI is pulling older ChatGPT models following GPT-5 launch - so bad news if you use GPT-4 or others at work". TechRadar. Retrieved 2025-08-09.
  42. ^ Robison, Kylie (7 August 2025). "OpenAI Finally Launched GPT-5. Here's Everything You Need to Know". Wired. ISSN 1059-1028. Retrieved 2025-08-07.
  43. ^ Roth, Emma (2025-08-08). "ChatGPT is bringing back 4o as an option because people missed it". The Verge. Retrieved 2025-08-09.
  44. ^ Li, Katherine. "OpenAI fans plead case to Sam Altman for GPT-4o's return". Business Insider. Retrieved 2025-08-09.
  45. ^ Whitwam, Ryan (2025-08-08). "ChatGPT users hate GPT-5's "overworked secretary" energy, miss their GPT-4o buddy". Ars Technica. Retrieved 2025-08-10.
  46. ^ Mauran, Cecily (2025-08-08). "Sam Altman: OpenAI will bring back GPT-4o after user backlash". Mashable. Retrieved 2025-08-10.
  47. ^ Nield, David (2025-08-09). "So many ChatGPT users have said they're missing the older GPT-4o model, OpenAI is going to bring it back". TechRadar. Retrieved 2025-08-09.
  48. ^ Field, Hayden (2025-08-13). "OpenAI will update GPT-5’s “personality” after user backlash.". The Verge. Retrieved 2025-08-13.
  49. ^ Varanasi, Lakshmi. "OpenAI is officially killing GPT-4o and users are freaking out (again)". Business Insider. Retrieved 2026-02-14.
  50. ^ Dellinger, A. J. (2026-02-13). "OpenAI Users Launch Movement to Save Most Sycophantic Version of ChatGPT". Gizmodo. Retrieved 2026-02-14.