List your company

ResearchFormerly WizWand

Visual Question Answering on SpecVQA

77.82Accuracy (English Descriptive L0)Gemini-3-Flash-Preview

Evaluation results

Method	Links
Gemini-3-Flash-Preview2026.04		77.82	77.59	80.47	79	78.72
Gemini-3-Pro-Preview2026.04		77.18	77.59	79.39	77.31	77.87
Gemini-2.5-Pro2026.04		76.45	74.86	77.58	75.71	76.15
Gemini-2.5-Flash2026.04		73.95	70.43	74.73	71.85	72.74
GPT-o4mini2026.04		71.44	68.36	72.37	69.96	70.54
GPT-5(high)2026.04		70.17	69.59	71.15	71.47	70.59
GPT-o32026.04		69.53	71	70.27	70.43	70.31
GPT-5(medium)2026.04		69.23	69.77	71.54	71	70.39
GPT-5(low)2026.04		68.25	69.68	70.66	69.87	69.61
GPT-5.12026.04		67.76	63.47	68.99	62.9	65.78
GPT-5.22026.04		67.76	63.28	68.99	61.58	65.4
Claude-4.5-Sonnet2026.04		61.48	55.18	63.15	57.82	59.41
Doubao-seed-1-6-flash2026.04		60.6	56.87	61.48	55.74	58.67
Claude-4-Sonnet2026.04		59.47	52.82	59.57	55.65	56.88
Qwen3-VL-8B-Thinking2026.04		58.64	53.48	58.05	55.08	56.31
Doubao-seed-1-6-2506152026.04		57.21	54.99	57.36	54.52	56.02
Qwen3-VL-8B-Instruct2026.04		57.21	42.56	59.27	47.27	51.58
Doubao-seed-1-62026.04		55.3	44.35	56.97	45.86	50.62
Qwen3-Omni-30B-A3B-Thinking2026.04		54.42	48.68	53.68	47.65	51.11
DeepSeek-VL22026.04		46.57	31.83	40.92	30.79	37.53