How many tracked prints does the MMStar family have?

MMStar leaderboard — the family board.

2 TRACKED VARIANTS · 1 FIRST-CLASS BOARD

One family, 2 tracked prints. Scoring configs and splits of the base board render as sections of this page, each under its own honest label.

● LIVE DATA53 MODELS ON THIS BOARDLEADER QWEN3.5 397B A17B AT 83.8

The board — MMStarTOP 53 OF 53 MEASURED · BEST SCORE PER MODEL, DEFAULT VARIANT

#ModelVendorScoreVLast seen

01Qwen3.5 397B A17BAlibaba83.82026-08-24

02Qwen3.6 PlusAlibaba83.32026-08-23

03Qwen3.5 122B A10BAlibaba82.92026-08-23

04Qwen3.5 35B A3BAlibaba81.92026-08-23

05Qwen3.6 27BAlibaba81.42026-08-24

06Qwen3.5 27BAlibaba812026-08-24

07Seed 1.8ByteDance79.92026-09-09

08Qwen3 VL 32B ReasoningAlibaba79.42026-08-23

09Gemini 2.5 ProGoogle79.22026-06-05

10Qwen3 VL 235B A22B ReasoningAlibaba78.72026-08-23

11Qwen3 VL 235B A22B InstructAlibaba78.42026-08-23

12Seed 1.5 VLByteDance77.92026-06-05

13Qwen3 VL 32B InstructAlibaba77.72026-08-23

14Step3 VL 10BStepFun77.62026-06-05

15Gemma 4 31BGoogle77.32026-08-24

16Qwen3 VL 30B A3B ReasoningAlibaba75.52026-08-23

17Qwen3.5 4BAlibaba75.32026-09-25

18Qwen3 VL Thinking (8B)Alibaba75.32026-08-23

19GLM-4.6V (106B-A12B)Z.ai75.32026-06-05

20EXAONE 4.5 33BLG AI74.92026-08-26

21GLM-4.6V-Flash (9B)Z.ai74.32026-06-05

22Qwen3 VL 4B (Reasoning)Alibaba73.22026-08-23

23Claude Opus 4.5Anthropic73.22026-08-24

24MiMo VL RL 2508 (7B)Xiaomi72.92026-06-05

25Qwen3 VL 30B A3B InstructAlibaba72.12026-08-23

26Qwen3 VL 8B InstructAlibaba70.92026-08-23

27Qwen2.5 VL 72BAlibaba70.82026-09-02

28InternVL-3.5 (8B)OpenGVLab69.82026-06-05

29Qwen3 VL 4B InstructAlibaba69.82026-08-23

30Qwen2.5 VL 32B InstructAlibaba69.52026-08-26

31Qwen2 VL 72B InstructAlibaba68.32026-08-25

32Qwen3.5 2BAlibaba67.92026-09-25

33Mage-VL-4BMicrosoft67.32026-07-26

34InternVL3_5-4BOpenGVLab65.52026-09-25

35Qwen2.5 Omni 7BAlibaba642026-08-23

36GPT-4oOpenAI63.92026-08-24

37LFM2.5-VL-3BLiquid AI63.32026-08-23

38Claude 3.5 SonnetAnthropic62.22026-08-24

39Gemma 4 E4BGoogle61.92026-09-25

40Phi 4 Multimodal InstructMicrosoft61.22026-07-26

41InternVL3-2BOpenGVLab61.12026-10-07

42Phi 4 R V 15BMicrosoft59.62026-07-26

43Gemma 4 E2BGoogle57.92026-09-10

44LFM2-VL-3BLiquid AI57.72026-08-22

45InternVL3_5-2BOpenGVLab57.72026-09-25

46Qwen2.5 VL 3BAlibaba56.12026-08-09

47North-Micro-Vision-InstructCohere51.82026-08-23

48LFM2.5-VL-1.6BLiquid AI50.72026-08-09

49InternVL3.5-1BOpenGVLab50.32026-08-09

50LFM2-VL-1.6BLiquid AI49.92026-10-07

51LFM2.5-VL-450M-ExtractLiquid AI432026-08-09

52LFM2-VL-450MLiquid AI40.92026-10-07

53MiMo Embodied 7BXiaomi22.42026-09-03

◆ IV — INDEPENDENTLY VERIFIED · ◆ AA — AGGREGATOR · ◆ VA — VENDOR · ◇ SA — SOURCE-ATTRIBUTED

Variants of this boardSCORING CONFIGS & SPLITS — NOT CITED STANDALONE, SO THEY LIVE HERE

MMStar (Acc)

VARIANT8 models · best score per model

#ModelVendorScoreVLast seen

01Qwen2.5 VL 72BAlibaba70.82026-06-05

02Kimi VL A3B ThinkingMoonshot70.42026-06-12

03Qwen2.5 VL 32B InstructAlibaba69.52026-06-05

04GPT-4oOpenAI64.72026-06-05

05Gemma 3 27BGoogle63.12026-06-05

06Kimi VL A3B InstructMoonshot61.72026-06-12

07Gemma 3 12BGoogle56.12026-06-05

08GPT-4o miniOpenAI54.82026-06-05

THE FAMILY BOARD · EVERY PRINT LABELED · EVERY SCORE SOURCE-ATTRIBUTED ON ITS OWN LIVE BOARDAll leaderboards The frontier board Models by vendor