// BENCHMARKS
Model Benchmarks 2025.
We tested four AI image models on 200 prompts across 10 categories. Here are the results.
Overall Rankings
| Model | Quality | Speed | Fidelity | Overall |
|---|---|---|---|---|
| Flux 1.1 Pro | 96/100 | 82/100 | 94/100 | #1 |
| DALL·E 3 | 89/100 | 71/100 | 93/100 | #2 |
| Stable Diffusion XL | 81/100 | 91/100 | 74/100 | #3 |
| Flux Schnell | 77/100 | 97/100 | 79/100 | #4 |
Category Breakdown: Photorealism
Flux 1.1 Pro dominates photorealistic outputs — portraits, landscapes, product shots. DALL·E 3 is close behind. Both significantly outperform SDXL for photo-quality results.
Category Breakdown: Illustration & Art
DALL·E 3 takes the top spot here, with exceptional compositional intelligence for illustrated work. Flux 1.1 Pro is a strong second. SDXL excels for anime and stylized art specifically.
Category Breakdown: Speed
| Model | Avg. Time | P90 Time |
|---|---|---|
| Flux Schnell | 2.1s | 3.4s |
| Stable Diffusion XL | 4.2s | 6.1s |
| Flux 1.1 Pro | 6.8s | 9.2s |
| DALL·E 3 | 11.4s | 16.3s |
Conclusion
For maximum quality: Flux 1.1 Pro. For creative work: DALL·E 3. For speed: Flux Schnell. All available free on TextToImg.