GPT-Image-2.5 (Sunburst) vs Nano Banana 3 Pro
A direct, empirical evaluation of GPT-Image-2.5 (Sunburst) (OpenAI) against Nano Banana 3 Pro (Google DeepMind).
GPT-Image-2.5 (Sunburst)
Developed by OpenAI · Proprietary Multimodal DiT parameters
High-fidelity typographic rendering, raytracing reflection accuracy, sketch-to-image conditioning, and localized comment inpainting.
Nano Banana 3 Pro
Developed by Google DeepMind · 2.8 Billion Parameters parameters
Zero-latency mobile on-device reasoning, 82ms TTFT, offline audio-visual tool execution, and local Android system integration.
| Metric | GPT-Image-2.5 (Sunburst) | Nano Banana 3 Pro |
|---|---|---|
| SWE-bench Verified | 62.4% | 58.6% |
| MMLU-Pro | 89.5% | 74.2% |
| MATH-500 | 91.2% | 82.5% |
| Context Window | Spatial Vector Token Canvas (1792×1024) | 131,072 tokens |
| Input Pricing (per 1M) | $2.50 | $0.00 |
| Output Pricing (per 1M) | $10.00 | $0.00 |