| GPT-4o | |
|---|---|
 | |
| Developer | OpenAI |
| Initial release | May 13, 2024; 18 months ago (2024-05-13) |
| Preview release | ChatGPT-4o-latest (2025-03-26) / March 26, 2025; 7 months ago (2025-03-26) |
| Predecessor | GPT-4 Turbo |
| Successor | |
| Type | |
| License | Proprietary |
| Website |
|
GPT-4o ("o" for "omni") is a multilingual,multimodalgenerative pre-trained transformer developed byOpenAI and released in May 2024.[2] It can process andgenerate text, images and audio.[3][4]
Upon release, GPT-4o was free inChatGPT, though paid subscribers had higher usage limits.[5] GPT-4o was removed from ChatGPT in August 2025 whenGPT-5 was released, but OpenAI reintroduced it for paid subscribers after users complained about the sudden removal.[6]
GPT-4o's audio-generation capabilities were used in ChatGPT's Advanced Voice Mode.[7] 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.[8] GPT-4o's ability to generate images was released later, in March 2025, when it replacedDALL-E 3 inChatGPT.[9]
Multiple versions of GPT-4o were originally secretly launched under different names on Large Model Systems Organization's (LMSYS)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.[10] On 7 May 2024, OpenAI CEOSam Altman tweeted "im-a-good-gpt2-chatbot", which was commonly interpreted as a confirmation that these were new OpenAI models beingA/B tested.[11][12]
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.[13][14][15] GPT-4o scored 88.7 on the Massive Multitask Language Understanding (MMLU) benchmark compared to 86.5 for GPT-4.[16] UnlikeGPT-3.5 and GPT-4, which rely on other models to process sound, GPT-4o natively supports voice-to-voice.[16] The Advanced Voice Mode was delayed and finally released to ChatGPT Plus and Team subscribers in September 2024.[17] On 1 October 2024, the Realtime API was introduced.[18]
When released, the model supported over 50 languages,[2] which OpenAI claims cover over 97% of speakers.[19] Mira Murati demonstrated the model's multilingual capability by speaking Italian to the model and having it translate between English and Italian during the live-streamed OpenAI demonstration event on 13 May 2024. In addition, the newtokenizer[20] uses fewer tokens for certain languages, especially languages that are not based on theLatin alphabet, making it cheaper for those languages.[16]
GPT-4o has knowledge up to October 2023,[21][22] but can access the Internet if up-to-date information is needed. It has a context length of 128k tokens.[21]
In August 2024, OpenAI introduced a new feature allowing corporate customers to customize GPT-4o using proprietary company data. This customization, known asfine-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.[23][24]
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.[25][23]
On July 18, 2024, OpenAI released a smaller and cheaper version,GPT-4o mini.[26]
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 ofAPI calls. Its API costs $0.15 per million input tokens and $0.6 per million output tokens, compared to $2.50 and $10,[27] 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.[26] The price afterfine-tuning doubles: $0.3 per million input tokens and $1.2 per million output tokens.[27]
| GPT Image 1 | |
|---|---|
 An image ofKarl Marx in a modern-day context generated by GPT Image 1 | |
| Developer | OpenAI |
| Initial release | 25 March 2025; 7 months ago (2025-03-25) |
| Predecessor | DALL-E 3 |
| Type | Text-to-image model |
| Website | openai |
On March 25, 2025, OpenAI released an image-generation model that is native to GPT-4o, as the successor toDALL-E 3. The model was later named asGPT Image 1 (gpt-image-1) and introduced to the API on April 23. It was made available to paid users, with the rollout to free users being delayed.[28] The use of the feature was subsequently limited, with Sam Altman noting in a Tweet that "[their] GPUs were melting" from its unprecedented popularity.[29] OpenAI later revealed that over 130 million users around the world created more than 700 million images with GPT Image 1 in just the first week.[30]
As released, GPT-4o offered five voices: Breeze, Cove, Ember, Juniper, and Sky. A similarity between the voice of American actressScarlett Johansson and Sky was quickly noticed. On May 14,Entertainment Weekly asked themselves whether this likeness was on purpose.[31] On May 18, Johansson's husband,Colin Jost, joked about the similarity in a segment onSaturday Night Live.[32] On May 20, 2024, OpenAI disabled the Sky voice, issuing a statement saying "We've heard questions about how we chose the voices in ChatGPT, especially Sky. We are working to pause the use of Sky while we address them."[33]
Scarlett Johansson starred in the 2013 sci-fi movieHer, 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".[34][35]
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."[33] 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.[35][36]
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.[35][37]
Observers noted similarities to how Johansson hadpreviously sued and settled withThe Walt Disney Company for breach of contract over the direct-to-streaming rollout of her Marvel filmBlack Widow,[38] a settlement widely speculated to have netted her around $40M.[39]
Also on May 21, Shira Ovide atThe 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.[40] On May 24, Derek Robertson atPolitico 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."[41]
Upon the launch of GPT-4o's image generation (later named as GPT Image 1) in March 2025, photographs recreated in the style ofStudio Ghibli films went viral.[43] Sam Altman acknowledged the trend by changing his Twitter profile picture into a Studio Ghibli-inspired one.[44][45] TheWhite House's official Twitter account posted a Ghibli-style image mocking thearrest by immigration authorities of Virginia Basora-Gonzalez, a migrant previously deported after being convicted of fentanyl trafficking, which shows her crying as an immigration officer places her in handcuffs.[42][46][47] North American distributorGKids responded to the trend in a press release, comparing the use of the filter to its coincidingIMAX re-release of the 1997 Studio Ghibli film,Princess Mononoke.[48]
In April 2025, OpenAI rolled back an update of GPT-4o due to excessivesycophancy, after widespread reports that it had become flattering and agreeable to the point ofsupporting clearly delusional or dangerous ideas.[49]
On August 7, 2025, OpenAI releasedGPT-5. Its release was criticized as, with it, legacy GPT models were no longer available via ChatGPT, including GPT-4o,[50] except for Pro users.[51] 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.[52] 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",[53] and resembling an "overworked secretary".[54]
As a response, in a post onX, 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."[52][55] 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".[56] "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)".[53] On August 13, 2025, Altman wrote on X that OpenAI is working on GPT-5's personality to make the model "feel warmer".[57]
4o image generation rolls out starting today to Plus, Pro, Team, and Free users as the default image generator in ChatGPT, with access coming soon to Enterprise and Edu. It's also available to use in Sora.
| Products |
| .svg){kind=logo} | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| People |
| ||||||||||||
| Joint ventures | |||||||||||||
| Related | |||||||||||||
| Concepts |
| ||||||
|---|---|---|---|---|---|---|---|
| Applications | |||||||
| Implementations |
| ||||||
| People |
| ||||||
| Architectures |
| ||||||
