List your company

Spatial Visual Question Answering on STVQA 7K (val)

93AccuracySpatialThinker-30B
49.52860.81472.183.386Nov 10, 2025

Evaluation results

MethodLinks
SpatialThinker-30BEvaluation Setting=Zero-shot, Decoding=Greedy, Input Modality=RGB, Training Dataset=STVQA-7K2025.1193
SpatialThinker-7BEvaluation Setting=Zero-shot, Decoding=Greedy, Input Modality=RGB, Training Dataset=STVQA-7K2025.1192.8
SpatialThinker-3BEvaluation Setting=Zero-shot, Decoding=Greedy, Input Modality=RGB, Training Dataset=STVQA-7K2025.1192.5
GPT-5-0807Evaluation Setting=Zero-shot, Decoding=Greedy, Input Modality=RGB2025.1189.7
Qwen2.5-VL-7B + Vanilla GRPOEvaluation Setting=Zero-shot, Decoding=Greedy, Input Modality=RGB, Training Dataset=STVQA-7K2025.1187.1
Qwen2.5-VL-3B + Vanilla GRPOEvaluation Setting=Zero-shot, Decoding=Greedy, Input Modality=RGB, Training Dataset=STVQA-7K2025.1186.7
Qwen2.5-VL-3B + SFTEvaluation Setting=Zero-shot, Decoding=Greedy, Input Modality=RGB, Training Dataset=STVQA-7K2025.1185.6
Qwen2.5-VL-7B + SFTEvaluation Setting=Zero-shot, Decoding=Greedy, Input Modality=RGB, Training Dataset=STVQA-7K2025.1184.5
Qwen3-VL-30BEvaluation Setting=Zero-shot, Decoding=Greedy, Input Modality=RGB2025.1183.7
Claude-4-Sonnet-0514Evaluation Setting=Zero-shot, Decoding=Greedy, Input Modality=RGB2025.1180.5
Qwen2.5-VL-7BEvaluation Setting=Zero-shot, Decoding=Greedy, Input Modality=RGB2025.1177.5
GPT-4o-0513Evaluation Setting=Zero-shot, Decoding=Greedy, Input Modality=RGB2025.1177
VLAA-Thinker-7BEvaluation Setting=Zero-shot, Decoding=Greedy, Input Modality=RGB2025.1176.2
SpaceThinkerEvaluation Setting=Zero-shot, Decoding=Greedy, Input Modality=RGB2025.1175.4
Visionary-R1Evaluation Setting=Zero-shot, Decoding=Greedy, Input Modality=RGB2025.1174.4
Qwen2.5-VL-3BEvaluation Setting=Zero-shot, Decoding=Greedy, Input Modality=RGB2025.1174.3
SpatialReasonerEvaluation Setting=Zero-shot, Decoding=Greedy, Input Modality=RGB2025.1174
SpaceOmEvaluation Setting=Zero-shot, Decoding=Greedy, Input Modality=RGB2025.1166
SATORI-R1Evaluation Setting=Zero-shot, Decoding=Greedy, Input Modality=RGB2025.1151.2