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

Chart Question Answering on ChartQA (Accuracy)

90.49Accuracy

MetaForge

Updated 18d ago

Evaluation Results

Method	Links
MetaForge 2026.06		90.49
MetaForge 2026.06		90.18
BF16 2026.06		89.44
SigLIP2-g 2026.07		89.4
RADIO1D 2026.07		89.2
RADIO1D 2026.07		89.2
C-RADIOv4-H 2026.07		89.1
RADIO1D 2026.07		89.1
SigLIP2-SO400m 2026.07		88.8
RADIO1D 2026.07		88.6
RADIO1D 2026.07		87.9
FP16 (Baseline) 2025.09		87.3
MODE 2026.06		87.26
InternVL3 2026.06		86.6
LLaVA-OneVision-1.5 2026.06		86.5
GPT-5-high 2026.06		86.2
RADIO1D 2026.07		85.8
VEQ-ME 2026.06		85.64
BF16 2026.06		85.2
AWQ 2025.09		84.5
MC-MoE 2026.06		84.26
GPTQ 2025.09		84.1
MoEQuant 2026.06		83.77
Claude-Opus-4.1 2026.06		83.3
MBQ 2026.06		83.13
Greedy 2026.04		83.12
MODE 2026.06		83
Speed-Q 2026.06		82.97
OmniQuant 2025.09		82.9
RADIO1D 2026.07		82.8
Greedy 2026.04		82.13
GPTQ 2026.06		82.04
Top-p (nucleus) 2026.04		81.13
VEQ-ME 2026.06		81.12
ε-Sampling 2026.04		80.16
η-Sampling 2026.04		80.16
Top-p (nucleus) 2026.04		80
MC-MoE 2026.06		79.94
ε-Sampling 2026.04		79.91
Top-p (nucleus) 2026.04		79.69
η-Sampling 2026.04		79.66
Min-p 2026.04		79.62
Top-k 2026.04		79.61
Temp' Only 2026.04		79.58
Temp’ Only 2026.04		78.96
Top-k 2026.04		78.96
Top-p (nucleus) 2026.04		78.93
η-Sampling 2026.04		78.73
SMMD 2026.06		78.38
ε-Sampling 2026.04		78.18
Temp’ Only 2026.04		77.97
Top-k 2026.04		77.92
NTIL 2026.06		77.08
CE 2026.06		77.02
NTL 2026.06		76.92
ε-Sampling 2026.04		76.53
η-Sampling 2026.04		76.46
Temp' Only 2026.04		76.14
Top-k 2026.04		76.11
Min-p 2026.04		76.09
GCE 2026.06		75.87
MODE 2026.06		75.2
CoLT 2026.06		74.7
SMMD 2026.06		73.21
Speed-Q 2026.06		73.2
MODE 2026.06		73.11
NTIL 2026.06		72.95
GCE 2026.06		72.32
SMMD 2026.06		72.21
NTL 2026.06		72.01
NTIL 2026.06		71.95
CE 2026.06		71.64
Min-p 2026.04		71.56
Min-p 2026.04		71.51
CE 2026.06		71.44
Min-p 2026.04		71.41
MoEQuant 2026.06		71.4
GCE 2026.06		70.44
Qwen2.5-VL-3B 2025.10		70.4
RECAP 2025.10		70.4
Qwen2.5-VL-3B-Instruct-GRPO-deepmath 2025.10		70.24
ε-Sampling 2026.04		69.79
Qwen2.5-VL-3B 2025.10		69.68
NTL 2026.06		69.66
LUQ 2025.09		68.6
RADIO1D 2026.07		68.2
MBQ 2026.06		68.16
vision-grpo-qwen-2.5-vl-3b 2025.10		67.8
Top-p (nucleus) 2026.04		67.6
LVR 2026.06		67.2
VEQ-ME 2026.06		67.16
SIM-CoT 2026.06		66.7
MC-MoE 2026.06		66.49
η-Sampling 2026.04		66.2
LaViDa-O + RL (Global + Factorized) 2026.06		66.16
GPTQ 2026.06		66.16
Qwen3-VL (Textual reasoning) 2026.06		65.1
GCE 2026.06		64.91
SMMD 2026.06		64.91
CE 2026.06		64.86

Showing 100 of 165 rows