The newest image generation models from OpenAI are now available globally for users of ChatGPT, ChatGPT Work, and Codex across desktop, mobile, and web platforms. This update introduces two API models, GPT-Image-2.5 Flare and GPT-Image-2.5 Sunburst, which provide higher fidelity and more precise editing controls. Flare delivers image generation with 50% lower latency than the previous GPT-Image 2 model, while Sunburst optimizes for high-precision creative work with longer generation times.
Sunburst ranks #1 and Flare ranks #2 in the Text-to-Image Arena benchmarks for image generation and editing categories. The release also adds a Sketch feature for drawing visual references and allows users to share prompts for others to remix. Third-party platform Venice integrated both Flare and Sunburst on September 8 to provide users with anonymous access to the models.
Key sources
- SOURCE@askvenice“Available anonymously.”x.com
- SOURCE@openai“two new models in the API: GPT-Image-2.5 Flare brings the same quality, editing, and speed improvements”x.com
- SUPPORT@openai“Comment-based edits to change only what you want”x.com
- SUPPORT@openaidevs“Flare delivers higher-quality images at 50% lower latency than GPT-Image-2”x.com
- SUPPORT@openaidevs“Change a jacket while keeping the subject in the same position”x.com
- SUPPORT@arena“Multi-Image Edit: +81 | +47 pts”x.com
- SUPPORT@sama“I don't think it can solve super difficult math problems”x.com
- SOURCEhuggingnewshuggingnews.com