List your company

ResearchFormerly WizWand

General multimodal reasoning

Benchmarks

DatasetSOTA methodMetricTrendResultsLast updated

Qwen2.5-VL-7B + SRPO

78.62General Score

21

May 11, 2026

Qwen2.5-VL-7B + SRPO

50.56Score

21

May 11, 2026

Qwen2.5-VL-7B + SRPO

38.09Score

21

May 11, 2026

General Multimodal Reasoning Suite MMK12, Geo3K, MathVerse, MathVista, We-Math

76.21MMK12 Accuracy

18

Jun 8, 2026

45.7Accuracy

14

Mar 4, 2026

Qwen2.5-VL-72B-IT

70.8Accuracy

13

Mar 4, 2026

MiMo-VL-7B-OpenMMReasoner

78.21Pass@1 Accuracy

11

Feb 26, 2026

Qwen3-VL-8B-DeepVision

71.33Pass@1 Accuracy

11

Feb 26, 2026

General Benchmarks

57.8Top-1 Accuracy

6

Apr 2, 2026