Head-to-head comparison
GPT Image 2.5 Sunburst vs. Nano Banana 2 Lite
Compare shared capability tests, arena scores with documented retrieval dates, task-specific image pricing, and verified specifications. Research benchmarks, preference arenas, and our n = 1 samples remain separate.
- shared result series
- 10
- shared sample prompts
- 0
- latest arena retrieval
- 2026-09-19
Change image model selection
Specifications and cost
General specifications and task-specific pricing
A model can have one price for generating new images and another for editing. Prices therefore remain tied to the corresponding arena configuration.
| Attribute | GPT Image 2.5 Sunburst | Nano Banana 2 Lite |
|---|---|---|
| Provider | OpenAI | |
| Released | September 8, 2026 | June 2026 |
| Access | API | API, Web app |
| Lifecycle | No separate lifecycle notice | No separate lifecycle notice |
| Weights | Proprietary | Proprietary |
| Parameters | Not published | Not published |
| Maximum output | 3,840 × 2,160 pixels | 1K |
| Arena configuration | Sunburst | Gemini 3.1 Flash Lite Image |
| Text-to-image price | No arena entry | $33.6 per 1,000 images |
| Editing price | No arena entry | $33.6 per 1,000 images |
| Technical source | Model sourceVerified 2026-09-19 | Model sourceVerified 2026-08-15 |
Prompt adherence and specialist tests
Shared capability benchmarks
These exact checkpoints have not both completed the fixed Gradually sample protocol. Earlier checkpoints do not substitute for them. No exact shared result from the integrated research benchmarks is available for this pair.
No shared Gradually ranking yet
A score appears only after both exact checkpoints complete all four fixed prompts and the same blind evaluation.
Independent preference tests
Shared arena benchmarks
Artificial Analysis and Arena derive scores from blind comparisons. Raters see outputs for the same prompt and choose the image they prefer. Separate series cover text rendering, photorealism, portraits, art, commercial design, and image editing. Confidence intervals indicate how certain a measured lead is.
Key to the marks:better valuebetter, but not conclusivetie
| Benchmark | GPT Image 2.5 Sunburst | Nano Banana 2 Lite | Source |
|---|---|---|---|
| LMArena Text to Image: 3D modeling | Higher score1,404.714 Rank 1, 95% confidence interval: ±41.7, 273 ratings | 1,226.293 Rank 13, 95% confidence interval: ±14.58, 1,568 ratings | LMArenaRetrieved 2026-09-19 |
| LMArena Text to Image: Art | Higher score1,390.12 Rank 1, 95% confidence interval: ±31.11, 406 ratings | 1,255.576 Rank 12, 95% confidence interval: ±12.19, 2,303 ratings | LMArenaRetrieved 2026-09-19 |
| LMArena Text to Image: Cartoon | Higher score1,439.841 Rank 1, 95% confidence interval: ±20.54, 1,198 ratings | 1,257.93 Rank 13, 95% confidence interval: ±8.68, 6,253 ratings | LMArenaRetrieved 2026-09-19 |
| LMArena Text to Image: Commercial design | Higher score1,418.818 Rank 1, 95% confidence interval: ±20.95, 1,104 ratings | 1,250.509 Rank 12, 95% confidence interval: ±8.73, 5,899 ratings | LMArenaRetrieved 2026-09-19 |
| LMArena Text to Image: Overall | Higher score1,420.715 Rank 1, 95% confidence interval: ±13.08, 3,149 ratings | 1,250.402 Rank 13, 95% confidence interval: ±6.19, 16,371 ratings | LMArenaRetrieved 2026-09-19 |
| LMArena Text to Image: Photorealism | Higher score1,423.225 Rank 1, 95% confidence interval: ±20.32, 1,168 ratings | 1,260.141 Rank 11, 95% confidence interval: ±8.56, 6,745 ratings | LMArenaRetrieved 2026-09-19 |
| LMArena Text to Image: Portraits | Higher score1,450.124 Rank 2, 95% confidence interval: ±31.56, 509 ratings | 1,247.68 Rank 15, 95% confidence interval: ±10.67, 3,352 ratings | LMArenaRetrieved 2026-09-19 |
| LMArena Text to Image: Text rendering | Higher score1,481.069 Rank 1, 95% confidence interval: ±22.66, 1,125 ratings | 1,275.491 Rank 12, 95% confidence interval: ±9.1, 5,412 ratings | LMArenaRetrieved 2026-09-19 |
| LMArena Image Editing: Multi-image editing | Higher score1,535.121 Rank 1, 95% confidence interval: ±9.09, 6,150 ratings | 1,289.38 Rank 14, 95% confidence interval: ±5.21, 26,348 ratings | LMArenaRetrieved 2026-09-19 |
| LMArena Image Editing: Overall | Higher score1,520.443 Rank 1, 95% confidence interval: ±9.21, 6,704 ratings | 1,314.296 Rank 19, 95% confidence interval: ±4.92, 59,280 ratings | LMArenaRetrieved 2026-09-19 |
Key to the marks:better valuebetter, but not conclusivetie | |||
Visual comparison
Sample images
Our images complement preference data, but do not replace it. The prompt, request count, and selection rule match. Provider resolution, quality tier, randomness, and prompt processing remain visible limitations, so this is not a fully controlled laboratory test.
No shared documented run yet
The fixed test suite covers typography, product photography, character details, and an infographic. A prompt appears only after both models have a verifiable first run.
Missing image run for GPT Image 2.5 Sunburst
Missing prompts. Typography and layout, Product photography, Character and detail, Infographic.
This checkpoint was released after the archived test and has no Gradually run yet. GPT Image 2 results remain assigned to GPT Image 2.
Next step. Run the four fixed prompts once with the exact gpt-image-2.5-sunburst API ID, then evaluate the outputs blind.
DocumentationHow to read this comparison
An Elo score is not a verdict on creativity
A higher arena score means this specific configuration was preferred more often in blind comparisons.
Typography, character consistency, local execution, licensing, and price may matter more to your workload than overall visual appeal.
Prices cover 1,000 images at 1,024 × 1,024 pixels using settings documented by Artificial Analysis. Subscriptions and special rates are excluded.
More matchups
Related image model comparisons
Compare GPT Image 2.5 Sunburst and Nano Banana 2 Lite with other important image models under the same evidence rules. Each page shows only exact shared measurements and samples.
View the complete image model comparison