Share your thoughts, 1 month free Claude Pro on usSee more

Visual Question Answering on AI2D

91.81Accuracy

GR3D-8B

Updated 17d ago

Evaluation Results

Method	Links
GR3D-8B 2026.05		91.81
GR3D-8B 2026.05		91.54
NVILA-Lite-8B 2026.05		91.01
Gemini 2.5 Pro 2026.04		88.4
OpenVLThinkerV2 2026.04		87.5
Baseline 2026.01		87.3
VISE 2026.06		87.24
VisionZero-Chart 2026.06		87.16
VisionZero-CLEVR 2026.06		87.1
iReasoner 2026.06		87.08
EvoLMM 2026.06		87.02
PTQ 2026.01		86.8
VisPlay 2026.06		86.71
QAD 2026.01		86.7
VisionZero-RW 2026.06		86.69
Base 2026.06		86.63
QAT 2026.01		86.5
Qwen3-VL-8B 2026.07		86.3
EMOVA 2024.09		85.8
InternVL3 2025.10		85.2
OneThinker-8B 2026.04		85.2
Vanilla 2026.06		85.2
Qwen3-VL GRPO 2026.04		85.1
Qwen3-VL GDPO 2026.04		85.1
GPT-4o 2026.04		84.9
VisionZero 2026.04		84.8
GPT-4o 2024.09		84.6
InternVL2.5 2025.10		84.5
MM-Eureka-7B 2026.04		84.1
VISE 2026.06		84.1
VisionZero-Chart 2026.06		84.04
Qwen3-VL-4B 2026.04		84
VisionZero-CLEVR 2026.06		83.98
iReasoner 2026.06		83.96
Qwen2.5-VL 2025.10		83.9
EvoLMM 2026.06		83.89
PTP 2025.09		83.8
FLARE-X 8B 2025.04		83.6
VisPlay 2026.06		83.45
Qwen2.5VL 7B 2025.04		83.4
VisionZero-RW 2026.06		83.37
Base 2026.06		83.31
NEO 2025.10		83.1
MoVA 2024.04		83
Qwen2-VL 2025.10		83
Encoder-Based 2025.10		82.9
Vanilla 2026.06		82.6
LLaVA-OneVision-7B 2026.04		82.51
FP16 Baseline 2026.03		82.42
InternVL2-8B 2025.09		82.4
RADIOv2.5-H + Tiling 2024.12		82.4
Qwen3-VL-Instruct-8B 2026.04		82.3
VISE 2026.06		82.16
GSearch 2025.09		82.1
Vision-G1 2026.04		82.1
VTW 2025.09		81.8
OpenVLThinker-7B 2026.04		81.8
VisionZero-Chart 2026.06		81.76
EMOVA 2024.09		81.7
Qwen2.5-VL 3B (reported) 2025.12		81.6
Qwen2.5-VL 2025.10		81.6
Qwen3-VL-4B-Instruct 2026.02		81.6
LLaVA-OneVision 7B 2025.04		81.6
FLARE-L 8B 2025.04		81.6
GeminiProVision 2024.06		81.4
Qwen2.5VL 3B 2025.04		81.4
FP16 Baseline 2026.03		81.31
iReasoner 2026.06		81.22
PALI-X-55B 2023.08		81.2
PALI-X-55B 2024.04		81.2
FLARE-X 3B 2025.04		81.2
EvoLMM 2026.06		81.13
MG-LLAVA 2024.06		81.1
FastV 2025.09		81.1
Molmo 7B 2025.04		81
VisionZero-CLEVR 2026.06		80.81
VL-Rethinker-7B 2026.04		80.8
SpB2.0-VL-5B 2026.04		80.73
RTN 2026.03		80.51
Qwen2.5-VL-3B 2026.04		80.51
PDrop 2025.09		80.4
VisPlay 2026.06		80.34
VisionZero-RW 2026.06		80.31
Gemini Pro 1.5 2024.09		80.3
FP16 Baseline 2026.03		80.12
Qwen2.5-VL 3B (reproduced) 2025.12		80.1
NEO 2025.10		80.1
Base 2026.06		80.1
QIG 2026.03		79.73
MBQ 2026.03		79.66
QIG 2026.03		79.63
AWQ 2026.03		79.47
MBQ 2026.03		79.47
FLARE-L 3B 2025.04		79.4
VITA 2024.09		79.3
SigLIP SO400M + Tiling 2024.12		79.3
Eagle2 2025.07		79.3
QIG 2026.03		79.11
InstructVLA-Generalist 2025.07		79.1
RTN 2026.03		79.02

Showing 100 of 402 rows