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

Multimodal Logical Reasoning on LogicVista

77Accuracy

Qwen3.5-27B

Updated 7d ago

Evaluation Results

Method
Qwen3.5-27B 2026.04	77	-
EXAONE 4.5 33B 2026.04	73.8	-
Qwen3-VL-235B-A22B 2026.04	72.2	-
Qwen3-VL-32B 2026.04	70.9	-
GPT-5 mini 2026.04	70.3	-
ChatGPT-4o-latest 2025.09	64.4	-
GPT-4o 2025.06	64.4	-
GPT-4.1-20250414 2025.09	61.1	-
Qwen2.5VL-72B x Qwen3-32B 2025.09	60.6	-
QVQ-72B-Preview 2025.09	58.2	-
Claude3.7-Sonnet 2025.09	58.2	-
Claude-3.7-Sonnet 2025.06	58.2	-
Qwen2.5VL-7B x Qwen3-32B 2025.09	56.6	-
Metis 2026.04	56.2	-
InternVL3-78B 2025.09	55.9	-
Qwen2.5-VL-72B 2025.09	55.7	-
Qwen2.5-VL-72B 2025.06	55.7	-
Qwen3-VL-8B-Instruct 2026.04	54.9	-
GLM-4v-Plus-20250111 2025.09	54.4	-
Gemini-2.0-Pro 2025.06	53.2	-
GPT-4o 2025.09	52.8	-
InternVL3.5-8B 2025.09	52.6	-
Qwen3-VL-4B-Instruct-Logic-RL Teacher 2026.05	52.5	73.8
Athena-7B 2025.06	51.3	-
GPT-4o 2025.11	51.2	-
Qwen2.5VL-7B x Qwen3-4B 2025.09	50.8	-
InternVL2.5-78B-MPO 2025.06	50.8	-
Thyme 2025.11	49	-
Thyme 2026.04	49	-
InternVL2.5-78B 2025.06	49	-
DeepEyesV2 2025.11	48.7	-
DeepEyesV2 2026.04	48.7	-
Keye-VL-8B 2025.09	48.6	-
VLAA-Thinker-7B 2025.09	48.5	-
Qwen2.5-VL-7B 2025.09	47.9	-
Qwen2.5-VL-7B 2025.06	47.9	-
DeepEyes 2025.11	47.7	-
DeepEyes 2026.04	47.7	-
MM-Eureka 2025.11	46.3	-
MM-Eureka-7B 2026.04	46.3	-
Qwen-2.5-VL 2025.11	45.9	-
VLAA-Thinker 2025.11	45.9	-
Qwen-2.5-VL-7B-Instruct 2026.04	45.9	-
VLAA-Thinker-7B 2026.04	45.9	-
SAIL-VL2-8B 2025.09	45	-
OpenVLThinker-7B 2025.09	44.5	-
InternVL3 2025.11	44.1	-
InternVL3-8B 2026.04	44.1	-
ThinkLite 2025.11	42.7	-
VL-Rethinker 2025.11	42.7	-
ThinkLite-VL-7B 2026.04	42.7	-
VL-Rethinker-7B 2026.04	42.7	-
Uni-OPD 2026.05	42	69.8
VLM-R1-3B-Math 2025.09	40.5	-
OPD 2026.05	38	65.2
Uni-OPD 2026.05	36.8	65.2
MiniCPM-o-2.6-8B 2025.06	36	-
OPD 2026.05	35.3	61.6
InternVL2.5-8B 2025.06	33.6	-
LLaVA-OV 2025.11	33.3	-
LLaVA-OneVision 2026.04	33.3	-
Qwen3-VL-2B-Instruct Student 2026.05	32.4	59.1
Qwen2.5-VL-3B 2025.09	28.2	-