List your company

Visual Question Answering on Colon-Bench Unprompted

82.5AccuracyGemini 3 Pro
36.53248.46660.472.334Mar 26, 2026

Evaluation results

MethodLinks
Gemini 3 Pro2026.0382.5
Gemini 3 ProUnprompted=true2026.0382.5
Gemini 3 Flash2026.0376
Gemini 3 FlashUnprompted=true2026.0376
Seed 1.62026.0372
Seed 1.6Unprompted=true2026.0372
Gemini 3.1 Flash Lite2026.0367.7
Gemini 3.1 F. Li.Unprompted=true2026.0367.7
Qwen3.5 Plus2026.0360.5
Qwen3.5 PlusUnprompted=true2026.0360.5
Qwen3.5 397B2026.0360.1
Qwen3.5 397BUnprompted=true2026.0360.1
Qwen3-VL 235B2026.0356.6
Qwen3-VL 235BUnprompted=true2026.0356.6
GLM-4.6V2026.0353.8
GLM-4.6VUnprompted=true2026.0353.8
Molmo 2-8B2026.0353.4
Molmo 2-8BUnprompted=true2026.0353.4
Qwen3-VL Plus2026.0350.5
Gemini 2.5 Flash Lite2026.0346.8
Gemini 2.5 F. Li.Unprompted=true2026.0346.8
Seed 1.6 Flash2026.0345.4
Qwen-VL Max2026.0345.4
Seed 1.6 FlashUnprompted=true2026.0345.4
Qwen-VL MaxUnprompted=true2026.0345.4
Qwen3-VL 32B2026.0344.4
Qwen3-VL 32BUnprompted=true2026.0344.4
Qwen3-VL 8B2026.0338.3
Qwen3-VL 8BUnprompted=true2026.0338.3