How many tracked prints does the RealWorldQA family have?

RealWorldQA 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 QWEN3.8 FLASH NEXT AT 88.5

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

#ModelVendorScoreVLast seen

01Qwen3.8 Flash NextAlibaba88.52026-08-27

02Qwen3.8 Max PreviewAlibaba882026-08-23

03Qwen3.7 Plus PreviewAlibaba86.92026-08-26

04Seed 2.1 Pro PreviewByteDance86.72026-09-28

05Seed2.1ByteDance86.32026-09-28

06Qwen3.8 27BAlibaba85.92026-08-26

07Gemini 3.1 ProGoogle85.42026-09-28

08Qwen3.6 PlusAlibaba85.42026-08-23

09Qwen3.6 35B A3BAlibaba85.32026-08-24

10Qwen3.5 122B A10BAlibaba85.12026-08-23

11Qwen3.6 27BAlibaba84.12026-08-24

12Qwen3.5 35B A3BAlibaba84.12026-08-24

13Qwen3.5 397B A17BAlibaba83.92026-08-24

14Qwen3.5 27BAlibaba83.72026-08-24

15GPT-5.5OpenAI82.22026-09-28

16Qwen3 VL 235B A22B ReasoningAlibaba81.32026-08-23

17Qwen3 VL 235B A22B InstructAlibaba79.32026-08-23

18Qwen3 VL 32B InstructAlibaba792026-08-23

19Qwen3 VL 32B ReasoningAlibaba78.42026-08-23

20Qwen2 VL 72B InstructAlibaba77.82026-08-24

21Qwen3 VL 30B A3B ReasoningAlibaba77.42026-08-23

22Claude Opus 4.5Anthropic772026-08-24

23Qwen3.5 4BAlibaba76.32026-09-25

24Claude Opus 4.7Anthropic75.62026-09-28

25GPT-4oOpenAI75.42026-08-24

26Claude Opus 4.6Anthropic73.92026-08-26

27Qwen2.5 VL 72BAlibaba73.92026-09-02

28Qwen3 VL 30B A3B InstructAlibaba73.72026-08-23

29Qwen3 VL Thinking (8B)Alibaba73.52026-08-23

30Qwen3 VL 4B (Reasoning)Alibaba73.22026-08-23

31LFM2.5-VL-3BLiquid AI73.12026-08-23

32Gemma 4 31BGoogle72.32026-08-24

33Gemma 4 26B A4BGoogle72.22026-08-24

34Qwen3 VL 8B InstructAlibaba71.52026-08-23

35Qwen3.5 2BAlibaba71.42026-09-25

36LFM2-VL-3BLiquid AI71.42026-08-22

37Qwen3 VL 4B InstructAlibaba70.92026-08-23

38Claude Sonnet 4.5Anthropic70.32026-08-24

39Qwen2.5 Omni 7BAlibaba70.32026-08-23

40InternVL3_5-4BOpenGVLab68.62026-09-25

41LFM2-VL-1.6BLiquid AI65.82026-10-07

42Qwen2.5 VL 3BAlibaba65.22026-08-09

43InternVL3-2BOpenGVLab65.12026-10-07

44LFM2.5-VL-1.6BLiquid AI64.82026-08-09

45Gemma 4 E4BGoogle64.32026-09-25

46North-Micro-Vision-InstructCohere62.22026-08-23

47InternVL3_5-2BOpenGVLab61.62026-09-25

48Claude 3.5 SonnetAnthropic60.12026-08-24

49Gemma 4 E2BGoogle602026-09-10

50LFM2.5-VL-450M-ExtractLiquid AI58.42026-08-09

51InternVL3.5-1BOpenGVLab57.12026-08-09

52LFM2-VL-450MLiquid AI52.02026-10-07

53MiMo Embodied 7BXiaomi28.52026-09-03

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

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

RealWorldQA (Acc)

VARIANT7 models · best score per model

#ModelVendorScoreVLast seen

01Qwen2.5 VL 72BAlibaba75.72026-06-05

02GPT-4oOpenAI75.42026-06-05

03Kimi VL A3B ThinkingMoonshot702026-06-12

04Kimi VL A3B InstructMoonshot68.12026-06-12

05GPT-4o miniOpenAI67.12026-06-05

06Gemma 3 27BGoogle62.52026-06-05

07Gemma 3 12BGoogle59.12026-06-05

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