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

Chart Understanding on ChartQA

91.2Accuracy

Socratic-Geo

Updated 13d ago

Evaluation Results

Method	Links
Socratic-Geo 2026.02		91.2
InternVL3-78B 2025.12		89.7
Qwen2.5-VL-72B 2025.12		89.5
Gemini 3-Pro 2026.02		89.44
DeepEyesV2 2025.11		88.4
LLaDA-o 2026.03		87.9
ERNIE 5.0 2026.02		87.8
CodeDance-7B 2025.12		87.5
InternVL3.5-30B-A3B 2026.04		87.4
Base Model 2026.02		87.3
Vanilla MoE 2026.04		87.26
MACS (Ours) 2026.04		87.14
Omni 2026.04		86.9
Qwen3-VL-30B-A3B-Instruct 2026.04		86.8
GPT-4o 2025.12		86.7
MACS (w/o Expanded) 2026.04		86.68
InternVL3 2025.11		86.6
Qwen2.5-VL-7B 2025.12		86.3
Qwen2.5-VL 2025.11		86.2
InternVL3-8B 2025.12		86.1
Thyme-VL-7B 2025.12		86.1
Thyme 2025.11		86.1
InternVL-3.5-4B 2026.02		86
GPT-4o 2024.08		85.7
Latent Denoising 2026.04		85.4
Vanilla MoE 2026.04		85.36
DenseMLLM-4B 2026.02		85.3
MACS (Ours) 2026.04		85.22
MACS (w/o Expanded) 2026.04		84.89
Qwen3-VL 2026.02		84.6
Qwen3-VL-4B 2026.02		84.6
Vanilla MoE 2026.04		84.14
Gemini 2.5-Pro 2026.02		84.08
Top-k Routing 2025.11		84.08
Qwen2.5-VL 3B (reported) 2025.12		84
MACS (Ours) 2026.04		83.98
CAI-MoE (Expanded) 2026.04		83.91
Gemini 2.5 Flash 2025.10		83.79
LLaVA-OneVision-72B 2024.08		83.7
Llava-OneVision-72B 2025.12		83.7
RISE 2026.05		83.64
Qwen2.5-VL 3B (reproduced) 2025.12		83.5
MACS (w/o Expanded) 2026.04		83.47
Gemini 2.5 Pro 2026.02		83.3
MC-MoE 2025.11		83.22
MoDES 2025.11		83.15
NAEE 2025.11		83.02
Baseline 2026.04		83
InternVL2.5-8B 2025.12		82.8
DiEP 2025.11		82.79
MoDES 2025.11		82.54
CASA⊕ Qwen2.5-VL 2025.12		82.4
Semantic-back-7B 2026.02		82.32
SAYO-InternVL-8B 2026.02		82.28
RISE 2026.05		82.28
RISE 2026.05		82.24
NoisyRollout-7B 2026.02		82.16
Kimi-VL-16B 2026.02		82.08
RISE 2026.05		82.08
Qwen2.5-VL 2025.10		82.03
Sparse-LaViDa 2025.12		82
MoDES 2025.11		82
SAYO-Qwen-4B 2026.02		81.96
SAYO-Qwen-8B 2026.02		81.84
RISE 2026.05		81.84
InternVL3.5-14B 2026.02		81.76
RISE 2026.05		81.68
InternVL3.5-38B 2026.02		81.6
InsertionHe-2B 2025.12		81.6
CAI-MoE (Expanded) 2026.04		81.57
GeoFocus 2026.02		81.5
Base Model 2026.05		81.44
Top-k Routing 2025.11		81.38
GRPO 2026.02		81.3
InternVL3.5-30B-A3B 2026.02		81.2
CAI-MoE (Token Drop) 2026.04		81.04
GPT-5 2025.10		80.85
InternVL3.5 2025.10		80.44
Llava-OneVision-7B 2025.12		80.4
Qwen3-VL-4B 2026.02		80.12
LLaVA-OneVision-7B 2024.08		80
LaViDa-O 2025.12		80
LaViDa-O 2026.03		80
LLaVA-OV 2025.11		80
VideoLLaMA3 2025.12		79.8
Base Model 2026.05		79.48
CAI-MoE (Expanded) 2026.04		79.28
InternVL2.5 2025.12		79.2
CAI-MoE (Token Drop) 2026.04		79.13
InternVL3.5-8B 2026.02		79.12
DiEP 2025.11		79.12
Baseline 2026.02		79
Qwen3-VL-8B 2026.02		78.96
GPT-4V 2024.08		78.5
Qwen3-VL-30B-A3B 2026.02		78.42
LLaDA-V 2026.03		78.3
GPT-5 2026.02		78.24
MC-MoE 2025.11		78.24
Deepeyes-7B 2025.12		78.2
Qwen2.5-VL 2025.10		78.12

Showing 100 of 159 rows