List your company

Reasoning-aware Image Generation on RiseBench 1.0 (test)

0.77Instruction ReasoningGemini-3-pro-image-preview
0.204240.351120.4980.64488Jan 6, 2026

Evaluation results

MethodLinks
Gemini-3-pro-image-preview2026.010.770.8550.9440.4120.6110.480.3760.472
GPT-Image-12026.010.6280.8020.9490.3410.3220.370.1060.289
Qwen-Image-Edit + Unified ThinkerThinker Backbone=Qwen3-VL-8B2026.010.6190.7620.9050.3290.30.410.0940.289
Gemini-2.5-Flash-Image2026.010.6120.860.9130.2590.4780.370.1880.328
Seedream-4.02026.010.5890.6740.9120.1290.1220.110.0710.108
BAGEL + Unified ThinkerThinker Backbone=Qwen3-VL-8B2026.010.5870.7570.8090.1520.1770.20.0820.155
Qwen-Image-Edit + Unified ThinkerThinker Backbone=Qwen2.5-VL-7B2026.010.5860.7590.9010.2470.2220.380.0940.242
GPT-Image-1-mini2026.010.5410.7150.9370.2470.2890.330.0940.244
BAGEL + Unified ThinkerThinker Backbone=Qwen2.5-VL-7B2026.010.5330.7360.7810.1410.1770.180.0350.136
Gemini-2.0-Flash-pre2026.010.4990.6840.8490.1060.1330.110.0230.094
Gemini-2.0-Flash-exp2026.010.4890.6820.8270.0820.1550.230.0470.133
BAGELChain of Thought (CoT)=true2026.010.4590.7380.8010.0590.1780.210.0120.119
Qwen-Image-Edit2026.010.3720.6640.8690.0470.10.170.0240.089
BAGEL2026.010.3650.5350.730.0240.0560.140.0120.061
Ovis-U12026.010.3390.5270.7290.0120.0330.040.0240.028
Step1X-Edit2026.010.3030.1260.74900.0220.020.0350.019
FLUX.1-Kontext-Dev2026.010.260.7160.8520.0230.0550.130.0120.058
OmniGen2026.010.2510.4150.7350.0120.0100.0120.008
EMU22026.010.2260.3820.7830.0120.011000.005