List your company

Mathematical Reasoning on GeoQA (test)

75.07AccuracyVIRC-7B
23.340436.770250.263.6298Dec 16, 2025Dec 25, 2025Jan 4, 2026Jan 14, 2026Jan 23, 2026Feb 2, 2026Feb 12, 2026

Evaluation results

MethodLinks
VIRC-7B#Params=7B, Model Category=Open-Source Reasoning Model2025.1275.07
Qwen3-VL-8B-Instruct#Params=8B, Model Category=Open-Source General Model, tool=true2025.1267.37
VIRC-3B#Params=3B, Model Category=Open-Source Reasoning Model2025.1267.37
MINT-CoT-7B#Params=7B, Model Category=Open-Source Reasoning Model2025.1264.72
MM-Eureka#Params=7B, Model Category=Open-Source Reasoning Model2025.1262.86
Pixel Reasoner#Params=7B, Model Category=Open-Source Reasoning Model2025.1253.05
InternVL2.5-8B-MPO#Params=8B, Model Category=Open-Source General Model2025.1252.2
Claude 3.5 Sonnet#Params=-, Model Category=Closed-Source Model2025.1250.8
DeepEyes#Params=7B, Model Category=Open-Source Reasoning Model2025.1249.8
DAPO + AT-RLModel=Qwen2.5-VL-7B2026.0248.55
Hint-GRPO-Qwen2-VL-7B#Params=7B, Model Category=Open-Source Reasoning Model2025.1248.14
DAPOModel=Qwen2.5-VL-7B2026.0246.95
R1-VL-7B#Params=7B, Model Category=Open-Source Reasoning Model2025.1246.86
GSPOModel=Qwen2.5-VL-7B2026.0246.8
Mini-o3#Params=7B, Model Category=Open-Source Reasoning Model2025.1246.02
GRPOModel=Qwen2.5-VL-7B2026.0245.92
Hint-GRPO-Qwen2.5-VL-3B#Params=3B, Model Category=Open-Source Reasoning Model2025.1244.3
LLaVA-OV-Qwen2-7b-ov#Params=7B, Model Category=Open-Source General Model2025.1244.16
Zero-shotModel=Qwen2.5-VL-7B2026.0244.05
DAPO + AT-RLModel=Qwen2.5-VL-3B2026.0243.85
Qwen2.5-VL-7B-Instruct#Params=7B, Model Category=Open-Source Reasoning Model2025.1243.5
FT-RLModel=Qwen2.5-VL-7B2026.0243.2
DAPOModel=Qwen2.5-VL-3B2026.0241.7
GSPOModel=Qwen2.5-VL-3B2026.0241.65
GRPOModel=Qwen2.5-VL-3B2026.0240.65
GPT-4o#Params=-, Model Category=Closed-Source Model2025.1240.53
Qwen2.5-VL-3B-Instruct#Params=3B, Model Category=Open-Source Reasoning Model2025.1239.66
Zero-shotModel=Qwen2.5-VL-3B2026.0238.72
FT-RLModel=Qwen2.5-VL-3B2026.0236.2
InternVL2.5-8B#Params=8B, Model Category=Open-Source General Model2025.1235.01
DeepSeek-VL2#Params=4.5B, Model Category=Open-Source General Model2025.1225.33