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

High-Resolution Multimodal Reasoning on HR-Bench 4K

88.3Overall Score

TTSP

Updated 5d ago

Evaluation Results

Method	Links
TTSP 2026.04		88.3	96.8	79.8
Deepconf 2026.04		86.4	96	76.8
RTWI 2026.04		86.4	95.8	77
SC 2026.04		86.3	95.8	76.8
TRT 2026.04		86.2	95.3	77
TTSP 2026.04		85.8	97.3	74.3
CISC 2026.04		85.7	95.5	75.8
Self-Cer. 2026.04		85.6	95.3	75.8
Self-Ref. 2026.04		85.4	93.8	77
RTWI 2026.04		84.3	95.5	73
Self-Ref. 2026.04		83.9	95.3	72.5
Deepconf 2026.04		83.9	95.3	72.5
SC 2026.04		83.7	95	72.3
TRT 2026.04		83.7	95	72.3
CISC 2026.04		83.6	94.8	72.3
Self-Cer. 2026.04		82.7	93	72.3
Base 2026.04		81	93	69
Base 2026.04		80	90	70
Mini-o3† 2026.03		77.5	-	-
Thyme 2026.04		77	91	63
LFPC 2026.03		76.4	92.7	60
LFPC 2026.03		75.3	86.8	63.7
DeepEyes‡ 2026.03		75.1	91.3	59
DeepEyes 2026.04		75.1	91.3	59
Pixel Reasoner 2026.03		73.6	85.8	61.5
CoF-sft 2026.03		73	92.2	53.7
DeepEyes 2026.03		72.9	90.5	55.2
Pixel-Reasoner 2026.04		72.9	86	60.3
V-Reflection 2026.03		72.6	83.5	61.8
Deepeyes 2026.03		71.3	83.8	58.8
Mini-o3 2026.03		70.7	82	59.3
InternVL3 2026.03		70	78.8	61.3
Qwen2.5-VL-7B 2026.03		69.2	78.8	59.8
Mini-o3 2026.03		69	84	54
Qwen2.5-VL-7B 2026.03		68	80.3	55.8
CoF-sft 2026.03		66	75.5	56.5
Pixel Reasoner 2026.03		65.8	69.8	61.8
GPT-4o 2026.03		65	66.8	63.3
GPT-4o 2026.04		65	66.8	63.3
DeepEyes 2026.03		63.9	73	54.7