List your company
ResearchFormerly WizWand

Geometry problem solving on Geo3K

98.2Top-1 AccuracyDraw2Think BL
0.12825.58951.0576.511Sep 30, 2025Nov 7, 2025Dec 16, 2025Jan 24, 2026Mar 3, 2026Apr 11, 2026May 20, 2026
Updated 5mo ago

Evaluation results

MethodLinks
Draw2Think BLBackbone=gemini-3-flash-preview, Execution Environment=direct VLM, Protocol=Training-free2026.0598.2
Draw2Think CTBackbone=gemini-3-flash-preview, Execution Environment=executable environment, Protocol=Training-free2026.0595.3
Human expertIdentity=College students2026.0590.9
OpenAI-o3Protocol=Zero-shot, Reported by=TrustGeoGen [9]2026.0581
Gemini-2.5-ProProtocol=Zero-shot, Reported by=TrustGeoGen [9]2026.0580.7
GPT-5-ThinkingModel Category=Close-source Models2025.0979.9
Pi-GPSProtocol=Neural-symbolic, Input Format=Parsed diagram annotations, Mechanism=PGDP + o3-mini solver2026.0577.8
Gemini-2.5-ProModel Category=Close-source Models2025.0977.2
PGPSNet-v2Protocol=Neural-symbolic, Input Format=Parsed diagram annotations, Mechanism=neural program synthesis2026.0576.4
GeoDRLProtocol=Neural-symbolic, Input Format=Parsed diagram annotations, Mechanism=RL + formal solver2026.0568.4
GeoGen-SFTProtocol=Fine-tuned, Backbone=qwen2.5-vl-7b, Mechanism=SFT2026.0558.4
Gemini 2.0 FlashModel Category=Closed-source LMMs2026.0258.2
InterGPSProtocol=Neural-symbolic, Input Format=Parsed diagram annotations, Mechanism=formal solver2026.0557.5
GeoFocus-7BModel Category=Open-source Mathematical Geometry LMMs, Parameters=7B2026.0255.3
VAPO-Thinker-7BModel Category=Our Models2025.0951.3
VLAA-Thinker-7BModel Category=Open-source Models2025.0950.6
GeoFocus-3BModel Category=Open-source Mathematical Geometry LMMs, Parameters=3B2026.0250.4
GeoUniModel Category=Open-source Mathematical Geometry LMMs2026.0250
Vision-R1-7BModel Category=Open-source Models2025.0949
GeoGen-SFT-7BModel Category=Open-source Mathematical Geometry LMMs, Parameters=7B2026.0246.3
R1-OneVision-7BModel Category=Open-source Models2025.0946.1
InternVL2.5-8BModel Category=Open-source Models2025.0944.8
Qwen2.5-VL-7BProtocol=Zero-shot, Input Format=Raw images, Reported by=GeoGen [35]2026.0544.8
VAPO-Thinker-3BModel Category=Our Models2025.0944.2
Qwen2.5-VL-7BModel Category=Open-source LMMs, Parameters=7B2026.0239.4
GeoGen-SFT-3BModel Category=Open-source Mathematical Geometry LMMs, Parameters=3B2026.0238.6
Qwen2.5-VL-7BModel Category=Open-source Models2025.0938.5
G-LLaVA-13BModel Category=Open-source Mathematical Geometry LMMs, Parameters=13B2026.0235
GPT-4oModel Category=Closed-source LMMs2026.0234.6
InternVL-3-8BModel Category=Open-source LMMs, Parameters=8B2026.0233.9
GPT-4oProtocol=Zero-shot, Reported by=TrustGeoGen [9]2026.0531.5
Claude 3.7Model Category=Closed-source LMMs2026.0231.1
InternVL-3-2BModel Category=Open-source LMMs, Parameters=2B2026.0231.1
Gemma-3-4BModel Category=Open-source LMMs, Parameters=4B2026.0230.6
GeoSketch-RLProtocol=Fine-tuned, Backbone=qwen2.5-vl-7b, Mechanism=SFT+RL2026.0528.8
Qwen2.5-VL-3BModel Category=Open-source LMMs, Parameters=3B2026.0225.3
InternVL-2.5-8BModel Category=Open-source LMMs, Parameters=8B2026.0216.6
mPLUG-Owl2-7BModel Category=Open-source LMMs, Parameters=7B2026.0212.7
Deepseek-VL-7BModel Category=Open-source LMMs, Parameters=7B2026.025.2
Monkey-Chat-7BModel Category=Open-source LMMs, Parameters=7B2026.024.7
Chameleon-7BModel Category=Open-source LMMs, Parameters=7B2026.023.9