OpenAI challenges Google's Nano Banana with faster GPT Image 1.5

OpenAI launched its latest image generation model, GPT Image 1.5, on December 16, directly aiming for market leadership against Google’s viral Gemini models, Nano Banana and Nano Banana Pro. The new model, available to all ChatGPT users and via the API, brings significant speed and quality improvements, tightening the competition in the highly visible AI image space.

Images 1.5 launches today in ChatGPT and the API!

Much better images in tons of ways, faster, and new editing capability.

— Sam Altman (@sama) December 16, 2025

4x speed and precise edits

OpenAI engineered GPT Image 1.5 to generate images up to four times faster than its predecessor. This speed increase not only improves the user experience within ChatGPT but also makes the model more practical for developers building high-volume image workflows.

The company focused heavily on improving image editing precision. GPT Image 1.5 handles targeted changes, such as adding, removing, or blending elements, while preserving crucial image details like lighting, composition, and facial features. For instance, a user can change a character’s outfit or add a specific background without the model accidentally altering the subject’s face or the scene’s overall light source. This level of consistency is key for creative professionals and marketing teams.

GPT Image 1.5 also shows improved capability in text rendering. This has been a persistent weakness across many generative AI models, often resulting in garbled or illegible words within an image. OpenAI reports the new model better handles denser and smaller text, making it a stronger tool for creating product mockups, infographics , and advertising creatives that require accurate, legible text.

The race against Nano Banana

OpenAI's launch follows a strong market showing by Google's Gemini-powered image models. The original Nano Banana (formally Gemini 2.5 Flash Image) became a viral sensation in late August 2025, praised for its speed, consistency, and ability to perform multi-turn, conversational edits. Google later released the more powerful Nano Banana Pro (formally Gemini 3 Pro Image) in November, which introduced high-resolution 4K output, multi-image referencing, and enhanced real-world knowledge grounding through Google Search integration.

Google's models found strong commercial traction by excelling in areas like brand consistency and text accuracy, features GPT Image 1.5 now directly addresses. Nano Banana Pro, for example, allows users to upload up to 14 reference images to maintain consistent brand identity, logos, and color palettes across outputs. It also connects to Google Search, allowing it to generate factually accurate diagrams and infographics based on real-world data.

Early tests show OpenAI gaining ground

Initial, independent blind-user testing by LMArena suggests OpenAI's upgrade successfully closes the quality gap. In preliminary results released on the launch day, GPT Image 1.5 briefly claimed the top position in text-to-image generation. The model's companion, the new ChatGPT Image Latest model, also secured the top spot for image editing.

However, the competition remains fierce. While GPT Image 1.5 showed a lead in raw generation, the margin in image editing between GPT Image 1.5 and Nano Banana Pro remains extremely narrow, indicating a near parity in the models' ability to handle complex, iterative edits. The final market winner will depend on continued improvements, deep platform integration, and developer pricing. OpenAI made its new model more cost-effective for developers, dropping API prices for image inputs and outputs by 20% compared to the previous version.

This release signals an intensified competition for the fastest and most capable AI image model, positioning GPT Image 1.5 as the direct, high-performance counter to Google’s popular Nano Banana series.