AIVideoFree adds OpenAI GPT Image 2.6 with Sunburst and Flare models

New models arrive in Singapore

The Singapore-based platform AIVideoFree has officially integrated OpenAI’s GPT Image 2.6. This integration occurred only two days after the model was first released by OpenAI on November 4, 2026. The platform allows users to access these new tools alongside existing options like GPT Image 2.5 and Nano Banana 2.1.

The update introduces two specific versions of the new model: Sunburst and Flare. These models are available through OpenAI’s API and are designed to serve different user needs. By offering both, the platform caters to people who prioritize either artistic perfection or quick turnaround times.

AIVideoFree provides a single interface where users can access various video and image models, including Veo, Kling, and Seedance. New users on the site receive free credits to test these tools, and the platform ensures that the generated outputs do not contain watermarks.

Speed versus artistic quality

OpenAI designed the two versions of GPT Image 2.6 to balance different priorities for different types of work. The Sunburst model is specifically tuned for high output quality. It is built to maintain consistent details across repeated edits, which is helpful for users who need to refine a specific image over time.

In contrast, the Flare model is built for speed. According to OpenAI, this version is capable of returning a generated image in about six seconds. While it may not prioritize the same level of granular consistency as Sunburst, it serves users who need rapid iterations or high-volume generation.

OpenAI reports that the new model produces visuals at a rate roughly half as fast as the previous iteration. This jump in performance is a significant technical milestone for the model family. It allows for a more responsive experience for creators who are trying to visualize ideas in real time.

Testing the text rendering hurdle

One of the most difficult tasks for image generators is rendering accurate text inside a picture. To see how the new models handle this, AIVideoFree conducted a side-by-side test. They used a complex prompt involving a 3D-animated Norwegian kitten in a Christmas setting with specific French calligraphic lettering.

The prompt required the model to display the phrase "Merci pour ton inscription" at the top of the image. By running the same prompt through both versions of GPT Image 2.6 and other models like Grok Imagine, the platform could provide a clear comparison. This method helps users see exactly where a new model improves over older technology.

The goal of these tests is to provide practical data for the community. As the market becomes crowded with new releases, AIVideoFree aims to show how these models perform under identical settings. This helps users decide which tool is best for their specific creative requirements.

Leaderboard success and future plans

The new Sunburst model has already seen significant success in independent testing. On the day of its release, DesignArena reported that Sunburst moved to first place on its Image Arena leaderboard. This put it ahead of the previous top-performing model, the GPT Image 2.5 Sunburst.

AIVideoFree intends to keep incorporating prominent image and video tools as they become available. The platform plans to continue publishing comparison data to help users navigate the rapidly changing landscape of AI tools. This approach is intended to provide a transparent look at how different models handle the same prompts.

Roger, the CEO of AIVideoFree, noted that because major models are released every few weeks, side-by-side testing is the most practical way to see improvements. The platform will continue to expand its library of models to ensure users have access to the latest technology from various providers.