Skip to content

Benchmark

RealWorldQA

Vision benchmarks measure how well models understand images. Scores are percentages. Higher is better.

Models measured
68
Scores recorded
79
Sources
3
IQ weight
0.95

Score distribution

73.1%
Mean
73.9%
Median
12.4%
Std dev
59.1%
10th percentile
86.9%
90th percentile
32.7%
Min
32.7%88.5%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.95 for chat models, measured over 67 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Qwen 3.8 Flash Next· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim88.5%
2Qwen 3.8 Flash NextQwen3.8-Flash-Next release/model card · published setting· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.88.5%
3Qwen3.8 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim88.5%
4Qwen3.8 FlashQwen3.8 Omni Flash 官方发布表· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.88.5%
5Qwen3.8 Max· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim88%
6Qwen3.8 MaxQwen3.8-Max release vision/video table · published setting· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.88%
7Qwen3.8 Omni FlashQwen3.8 Omni Flash 官方发布表· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.87.7%
8Qwen3.7 Plus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim86.9%
9Qwen3.7 PlusQwen3.7-Plus release · published setting· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.86.9%
10Qwen3.7 PlusQwen3.8 Omni Flash 官方发布表· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.86.9%
11Doubao Seed 2.1 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim86.7%
12Doubao Seed 2.1 Turbo· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim86.3%
13Qwen 3.8 27B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim85.9%
14Qwen 3.8 27BQwen3.8 Omni Flash 官方发布表· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.85.9%
15Qwen 3.8 27BQwen3.8-27B model card · published setting· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.85.9%
16Qwen3.6 Plus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim85.4%
17Qwen3.6 PlusQwen official release table · thinking mode unless benchmark requires otherwise· Reported by Alibaba Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.85.4%
18Qwen3.6 35B A3B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim85.3%
19Qwen3.6 FlashQwen3.6-35B-A3B official model card · hosted API name Qwen3.6-Flash· Reported by Alibaba Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.85.3%
20Qwen3.5 122B-A10B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim85.1%
21Qwen3.5 35B A3B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim84.1%
22Qwen3.5 Omni PlusQwen3.5-Omni release · published setting· Reported by Qwen, Qwen Team· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.84.1%
23Qwen3.6 27B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim84.1%
24Qwen3.6 27BQwen official release table · thinking mode unless benchmark requires otherwise· Reported by Alibaba Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.84.1%
25Qwen3.5 397B A17BQwen3.5 release/model card · published setting· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.83.9%
26Qwen3.5 27B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim83.7%
27Seed 2.0 Mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim81.6%
28Qwen3 VL 235Bthinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.81.3%
29Qwen3 VL 235B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.79.3%
30Qwen3 VL 32B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.79%
31GPT 4.1· OpenVLM Leaderboard data78.7%
32InternVL3 78B· OpenVLM Leaderboard data78.4%
33Qwen3 VL 32B Thinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.78.4%
34Qwen2-VL-72B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
77.8%
35GPT-4.1 mini· OpenVLM Leaderboard data77.5%
36Qwen3.5 Omni FlashQwen3.5-Omni release · published setting· Reported by Qwen, Qwen Team· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.77.5%
37qwen3-vl-30bthinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.77.4%
38internvl3-38b· OpenVLM Leaderboard data75.8%
39Qwen2.5 VL 72B Instruct· OpenVLM Leaderboard data75.3%
40Claude Opus 4.6Qwen3.8 Omni Flash 官方发布表 · max effort· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.73.9%
41Molmo 72B· OpenVLM Leaderboard data73.7%
42qwen3-vl-30b· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.73.7%
43Qwen: Qwen3 VL 8B Thinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.73.5%
44Gemini 2.0 Flash· OpenVLM Leaderboard data72.3%
45Qwen3 VL 8B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.71.5%
46internvl3-8b· OpenVLM Leaderboard data71.4%
47Molmo 7B-D· OpenVLM Leaderboard data71.1%
48GLM 4 Vision Plus· OpenVLM Leaderboard data70.5%
49Qwen2.5-Omni 7B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim70.3%
50DeepSeek VL2· OpenVLM Leaderboard data
Values from 2 sources
70.1%

Showing 1–50 of 79 results.

Data credits: OpenVLM Leaderboard (OpenCompass / VLMEvalKit), Apache-2.0 (licence) · BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links. · Data by LLM Stats