Skip to content

Benchmark

LVBench

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

Models measured
35
Scores recorded
47
Sources
2
IQ weight
2.12

Score distribution

71.7%
Mean
75.4%
Median
12%
Std dev
58%
10th percentile
85.4%
90th percentile
40.4%
Min
40.4%91.7%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.12 for chat models, measured over 35 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Gemini 4 Argon· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim91.7%
2Gemini 4 Argonlong-video understanding · official launch-page result· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.91.7%
3Gemini 3.8 FlashGemini为agentic;跨模型设置见原卡· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.87.8%
4Gemini 3.8 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim87.1%
5Qwen3.8 MaxQwen3.8-Max release vision/video table · Qwen-MM-Plugins memory system· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.85.6%
6Gemini 3.7 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim85.4%
7Gemini 3.7 Flashno tools · 1,024 frames· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.85.4%
8GPT-5.6 SolGemini为agentic;跨模型设置见原卡· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.82.1%
9Qwen3.8 Max· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim81.8%
10Qwen3.8 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.81.8%
11Doubao Seed 2.1 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim78%
12Qwen3.8 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.76.9%
13Doubao Seed 2.1 Turbo· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim76.8%
14Qwen 3.8 Flash Next· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim76.6%
15Qwen 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.76.6%
16Qwen3.8 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim76.6%
17Qwen3.8 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.76.6%
18Qwen3.7 Plus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim76.2%
19Qwen3.7 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.76.2%
20Qwen3.7 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.76.2%
21Kimi K2.5setting not named by the source· Reported by Moonshot AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.75.9%
22Kimi K2.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim75.9%
23Qwen3.5 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.75.5%
24Claude Opus 5Gemini为agentic;跨模型设置见原卡· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.75.4%
25Qwen3.5 122B-A10B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim74.4%
26Qwen3.5 27B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim73.6%
27Doubao Seed 1.8· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim73%
28Qwen 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.72.4%
29Qwen3.5 35B A3B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim71.4%
30Qwen3.6 35B A3B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim71.4%
31Qwen3.6 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.71.4%
32Qwen3.5 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.71.2%
33Qwen3 VL 235B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.67.7%
34Seed 2.0 Mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim66.6%
35Qwen3.5 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.65.7%
36Qwen3 VL 32B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.63.8%
37Qwen3 VL 235Bthinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.63.6%
38Claude 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.63%
39Qwen3 VL 32B Thinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.62.6%
40qwen3-vl-30b· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.62.5%
41qwen3-vl-30bthinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.59.2%
42Qwen3 VL 8B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.58%
43Qwen: Qwen3 VL 8B Thinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.55.8%
44Qwen2.5 VL 72B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim47.3%
45Qwen2.5-VL 7B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim45.3%
46Nova Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim41.6%
47Nova Lite· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim40.4%

Data credits: 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