List your company
ResearchFormerly WizWand

Visual Question Answering on AI2D (test)

96.4AccuracyMolmoPoint-8B
14.999236.132157.26578.3979Oct 7, 2022May 16, 2023Dec 24, 2023Aug 2, 2024Mar 11, 2025Oct 19, 2025May 29, 2026
Updated 5mo ago

Evaluation results

MethodLinks
MolmoPoint-8BAccess=MolmoPoint2026.0396.4
Molmo-72BAccess=Open weights, data, code, Number of crops=36, Style Prompting=AI2D with transparent boxes2024.0996.3
Gemini 2.5 FlashAccess=API call only2026.0395.9
Molmo2-8BAccess=Fully open2026.0395.8
Molmo2-4BAccess=Fully open2026.0395.6
Claude-3.5 SonnetAccess=API2024.0994.7
Gemini 1.5 ProAccess=API2024.0994.4
Gemini 2.5 ProAccess=API call only2026.0394.3
GPT-4o-0513Access=API2024.0994.2
MolmoPoint-8B-O-7BAccess=Fully open2026.0393.7
Molmo-7B-DAccess=Open weights, data, code, Number of crops=36, Style Prompting=AI2D with transparent boxes2024.0993.2
PLM-8BAccess=Fully open2026.0392.7
Llama-3.2V-90B-InstructAccess=Open Weights2024.0992.3
Gemini 1.5 FlashAccess=API2024.0991.7
Claude Sonnet 4.5Access=API call only2026.0391.5
Llama-3.2V-11B-InstructAccess=Open Weights2024.0991.1
PLM-3BAccess=Fully open2026.0390.9
Molmo-7B-OAccess=Open weights, data, code, Number of crops=36, Style Prompting=AI2D with transparent boxes2024.0990.7
Zamba2-VL-7BLanguage-backbone scale=7–8B2026.0590.6
GPT-5Access=API call only2026.0389.5
Keye-VL-1.5-8BAccess=Open weights only2026.0389.5
GPT-4VAccess=API2024.0989.4
Claude-3 OpusAccess=API2024.0988.1
Qwen2-VL-72BAccess=Open Weights2024.0988.1
GLM-4.1V-9BAccess=Open weights only2026.0387.9
InternVL2-Llama-3-76BAccess=Open Weights2024.0987.6
Claude-3 HaikuAccess=API2024.0986.7
GPT-5 miniAccess=API call only2026.0386.7
MiniCPM-V-4.5-8BAccess=Open weights only2026.0386.5
MolmoE-1BAccess=Open weights, data, code, Number of crops=36, Style Prompting=AI2D with transparent boxes2024.0986.4
Zamba2-VL-2.7BLanguage-backbone scale=2–4B2026.0585.8
Qwen3-VL-8BAccess=Open weights only2026.0385.7
LLaVA OneVision-72BAccess=Open Weights + Data, distilled=true2024.0985.6
Eagle2.5-8BAccess=Open weights only2026.0384.5
Qwen3-VL-4BAccess=Open weights only2026.0384.1
InternVL3.5-8BAccess=Open weights only2026.0384
InternVL2-8BAccess=Open Weights2024.0983.8
Mamba-VL-2.8BLanguage-backbone scale=2–4B2026.0583.7
Qwen2-VL-7BAccess=Open Weights2024.0983
InternVL3.5-4BAccess=Open weights only2026.0382.6
SMoLA-PaLI-X_FT (Specialist)Base model=PaLI-X, Per-task LoRA tuning=true, LoRA rank=42023.1282.5
jina-vlm2025.1282
Pythia-VL-2.8BLanguage-backbone scale=2–4B2026.0581.6
Zamba2-VL-1.2BLanguage-backbone scale=~1B2026.0581.5
PaLI-XOCR pipeline input=true2023.0581.4
SMoLA-PaLI-XFTOCR pipeline input=false, Model Category=Generalist, Backbone=PaLI-X2023.1281.4
PaLI-X (Specialist)OCR pipeline input=true, Model Category=Specialist, Reference Citation=[7]2023.1281.4
SMoLA-PaLI-XFTOCR pipeline input=true, Model Category=Generalist, Backbone=PaLI-X2023.1281.4
PaLI-XBase model=PaLI-X2023.1281.4
SOTA [7]2023.1281.4
LLaVA OneVision-7BAccess=Open Weights + Data, distilled=true2024.0981.4
PaLI-XOCR pipeline input=false2023.0581.2
PaLI-X (Specialist)OCR pipeline input=false, Model Category=Specialist, Reference Citation=[7]2023.1281.2
Mamba-VL-1.4BLanguage-backbone scale=~1B2026.0580.9
Cambrian-1-34BAccess=Open Weights + Data, distilled=true2024.0979.7
Mamba-VL-0.8BLanguage-backbone scale=~1B2026.0579.3
Pythia-VL-1.4BLanguage-backbone scale=~1B2026.0579.3
Pixtral-12BAccess=Open Weights2024.0979
InternVL3.5-2BParameter Count=2B2025.1278.8
InternVL3-2BParameter Count=2B2025.1278.6
Phi3.5-Vision-4BAccess=Open Weights2024.0978.1
Pythia-VL-1BLanguage-backbone scale=~1B2026.0577.6
Qwen3-VL-2BParameter Count=2B2025.1276.9
CogVLMOCR pipeline input=false, Model Category=Generalist2023.1275.7
SMoLA-PaLI-3FTOCR pipeline input=false, Model Category=Generalist, Backbone=PaLI-32023.1275.6
SMoLA-PaLI-3FTOCR pipeline input=true, Model Category=Generalist, Backbone=PaLI-32023.1275.5
Qwen2-VL-2BParameter Count=2B2025.1274.7
xGen-MM-interleave-4BAccess=Open Weights + Data, distilled=true2024.0974.2
Cambrian-1-8BAccess=Open Weights + Data, distilled=true2024.0973
PaliGemma-mix-3BAccess=Open Weights2024.0972.3
Qwen-VLOCR pipeline input=false, Model Category=Generalist2023.1262.3
LLaVA-1.5-13BAccess=Open Weights + Data2024.0961.1
LLaVA-1.5-7BAccess=Open Weights + Data2024.0955.5
Method [27]OCR pipeline input=false2023.0542.1
Pix2Struct LargePretraining=Screenshot parsing, Pixel only=true2022.1042.1
Pix2Struct BasePretraining=Screenshot parsing, Pixel only=true2022.1040.9
Method [45]OCR pipeline input=true2023.0538.5
DQANtype=State of the art w/ pipelines2022.1038.5
DonutPretraining=OCR, Pixel only=true2022.1030.8
LAMMSetting=Zero-shot2023.0620.92
mPLUG-owlSetting=Zero-shot2023.0619.31
LLaVASetting=Zero-shot2023.0618.13