🔬 Tested by TokenRouter

Same prompt. Every model. Judge for yourself.

We run a fixed set of prompts against every model in the catalog and keep the outputs — so you can compare what they actually produce, not what the docs claim.

The prompt Generate a wide 16:9 landscape photo of a mountain range.
flux-2-flex black-forest-labs ✗ failed 13s
flux-2-max black-forest-labs ✗ failed 24s
flux-2-pro black-forest-labs ✗ failed 11s
flux-kontext-max black-forest-labs ✗ failed 10s
flux-kontext-pro black-forest-labs ✗ failed 8s
FLUX.1 Kontext [max] black-forest-labs ✗ failed 9s
FLUX.1 Kontext [pro] black-forest-labs ✗ failed 7s
FLUX1.1 [pro] black-forest-labs ✗ failed 4s
FLUX.2 [dev] black-forest-labs ✗ failed 5s
FLUX.2 [flex] black-forest-labs ✗ failed 10s
FLUX.2 [max] black-forest-labs ✗ failed 32s
seedream-4-0-250828 bytedance ✗ failed 7s
gemini-2.5-flash-image google 5s
gemini-3-pro-image google 16s
gemini-3.1-flash-image google 7s
gemini-3.1-flash-image google 7s $0.0906
gemini-3.1-flash-lite-image google ✓ passed 5s
OpenAI: GPT-5 Image openai 58s
OpenAI: GPT-5 Image Mini openai 44s
OpenAI: GPT-5.4 Image 2 openai 195s
gpt-image-1 openai ✗ failed 37s
gpt-image-1-mini openai ✗ failed 39s
gpt-image-1.5 openai ✗ failed 39s
gpt-image-1.5-2025-12-16 openai ✗ failed 35s
gpt-image-2 openai ✗ failed 30s
gpt-image-2-all openai ✗ failed 25s
gpt-image-2-vip openai ✗ failed 76s
Qwen Image qwen ✗ failed 11s
Juggernaut Lightning Flux by RunDiffusion rundiffusion ✗ failed 5s
Juggernaut Pro Flux by RunDiffusion 1.0.0 rundiffusion ✗ failed 4s
SD XL stabilityai ✗ failed 4s

Outputs generated 2026-08-26. Costs are measured from the provider's own billing where reported. Nothing here is simulated.