GPT Image 2.6 · Early access
Flare vs Sunburst: choosing the right GPT Image 2.6 model
The two models are tuned for different stages of the same workflow.
| Flare | Sunburst | |
|---|---|---|
| Core advantage | Speed for iteration | Detail and consistent revisions |
| Speed | At low quality, results can take ~6s; benchmark: 4.84s compared with 7.08s for Images 2.5. | Designed for detail rather than Flare-level speed. |
| Suggested quality | Low | Medium or high |
| Edit continuity | Well suited to trying ideas quickly. | Best consistency through medium- and high-quality edits. |
| Good fit | Fast drafts and early-stage concepts | Precise scenes and later-stage polish |
| Credits | 1 credit per 1K image | 4 per 1K image, 8 per 2K image |
Draft with Flare, finish with Sunburst: settle the direction at low cost, then render the approved version at higher quality.
Get product updates
Get GPT Image 2.6 product updates by email