# GPT Image 1.5

GPT Image 1.5 is an AI image generation model developed by OpenAI, released in 2025 as an update to GPT Image 1. It offers enhanced photorealism, text rendering, and editing capabilities through the ChatGPT interface and API.

GPT Image 1.5 is a [generative artificial intelligence](https://www.wikiprompt.org/wiki/generative-ai) model developed by [OpenAI](https://www.wikiprompt.org/wiki/openai) for image creation and editing. Released in 2025, it is the successor to GPT Image 1 and is integrated into ChatGPT and the OpenAI API, allowing users to generate and modify images from natural language prompts. The model builds on OpenAI's [large language model](https://www.wikiprompt.org/wiki/large-language-model) architecture, combining multimodal understanding with diffusion-based image synthesis.

GPT Image 1.5 was announced by OpenAI in early 2025, with a public release in March 2025. It became available to ChatGPT Plus, Pro, and Team subscribers, as well as through the API for developers. The model is also accessible via OpenAI's enterprise offerings, including Microsoft Azure's [Azure OpenAI Service](https://www.wikiprompt.org/wiki/azure).

## Capabilities

GPT Image 1.5 introduces several improvements over its predecessor, including higher resolution output (up to 2048x2048 pixels), better adherence to complex prompts, and more accurate rendering of text within images. The model supports multi-turn editing, allowing users to refine images through conversational instructions. It can also generate images with consistent characters and styles across multiple outputs, a feature aimed at creative professionals.

According to OpenAI's technical documentation, GPT Image 1.5 uses a hybrid approach that combines an autoregressive transformer for layout and composition with a diffusion decoder for pixel-level detail. This enables more coherent scenes and reduces common artifacts such as distorted hands or garbled text.

## Performance and Benchmarks

In internal evaluations, OpenAI reported that GPT Image 1.5 outperformed GPT Image 1 on the MMMU (Multimodal Multidisciplinary Understanding) benchmark and human preference tests for photorealism and prompt alignment. Independent benchmarks, such as those from the Artificial Analysis intelligence index, placed GPT Image 1.5 among the top image generation models in 2025, rivaling systems from [Google DeepMind](https://www.wikiprompt.org/wiki/google-deepmind) and other labs.

The model also demonstrated improved capabilities in generating images with fine-grained details, such as small text, logos, and scientific diagrams. OpenAI noted that GPT Image 1.5 reduced errors in rendering long passages of text by approximately 40% compared to the previous version.

## API and Pricing

GPT Image 1.5 is available through the OpenAI API with a tiered pricing structure. As of the release, the model was priced at $0.02 per image for standard resolution (1024x1024) and $0.08 per image for high resolution (2048x2048). The API supports parameters for quality, size, and style, and integrates with the existing chat completions endpoint.

Developers can use GPT Image 1.5 for applications ranging from marketing content generation to design prototyping. OpenAI also offers fine-tuning options for enterprise customers, allowing customization on specific datasets, though this feature is limited to select partners.

## Safety and Limitations

OpenAI implemented several safety measures for GPT Image 1.5, including content filters to prevent the generation of harmful or deceptive imagery. The model includes a C2PA watermark to indicate AI-generated content, and OpenAI has restricted the generation of realistic faces of public figures without consent. Despite these measures, the model can still produce biased or inaccurate representations, and OpenAI advises users to review outputs for factual correctness.

As with other [neural network](https://www.wikiprompt.org/wiki/neural-network) models, GPT Image 1.5 has limitations in understanding abstract concepts and may struggle with highly specific or ambiguous prompts. The model's training data includes publicly available images and text from the internet, which can introduce cultural biases.

## Reception

GPT Image 1.5 received positive reviews from technology media and creative communities for its improved text rendering and editing capabilities. Critics noted that while it does not fully replace professional design tools, it significantly lowers the barrier for rapid visual prototyping. Some concerns were raised about the potential for misuse in creating misleading images, prompting OpenAI to update its usage policies.

By mid-2025, GPT Image 1.5 was widely adopted in industries such as advertising, e-commerce, and education. It also became a common tool in academic research for generating illustrative figures and data visualizations.

## See Also

- [Generative AI](https://www.wikiprompt.org/wiki/generative-ai)
- [OpenAI](https://www.wikiprompt.org/wiki/openai)
- [Large language model](https://www.wikiprompt.org/wiki/large-language-model)
- [Artificial intelligence](https://www.wikiprompt.org/wiki/artificial-intelligence)

---
Source: https://www.wikiprompt.org/wiki/gpt-image-1-5
License: CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/)
Last updated: 2026-09-13T18:55:54.898223+00:00