OCR-VQA
Models answer questions about book covers by reading visible text. The benchmark compares answers with facts such as a book’s title or author.
- Made by
- Anand Mishra and colleagues
- Released
- September 2019
- Kind
- Documents, charts and text in images
- Level
- School
- Scored
- Checked automatically
- Answers
- Short answer
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Single turn
- Items from
- Real work and use
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Vision
- Strongest model tested
- GPT-4.1 #138 on the capability index
- Last model added
- April 2025
- Scores refreshed
- November 20th, 2026
Kimi-VL-A3B-Instruct leads with 82% among models released by April 2025 Kimi-VL-A3B-Instruct leads with 82% among models released by April 2025
Best score per entry from OpenCompass’s leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.
The newest model with a score here came out in April 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 82.0 Kimi-VL-A3B-Instruct
- 80.9 MUG-U-7B
- 75.1
Ovis1.6-Gemma2-27B
- 73.7 Qwen2-VL-72B-Instruct
- 73.5 Taiyi
- 71.2
Ovis1.5-Llama3-8B
- 71.2
Ovis1.6-Gemma2-9B
- 70.3 H2OVL-2B
- 68.2
Cambrian-1-34B
- 67.5 Qwen-VL
- 67.0 Monkey
- 67.0
DeepSeek-VL-7B-Chat
- 66.7
Ovis1.6-Llama3.2-3B
- 66.3 CogVLM-17B-Chat
- 66.3
LLaVA-v1.6-34B
- 66.0
Cambrian-1-8B
- 65.3 llava-v1.6-vicuna-13b
- 65.2
mPLUG-Owl2
- 65.0
Eagle-X5-13B
- 64.9 CogVLM2-19B-Chat
- 64.7
LLaVA-OneVision 7B (SI)
- 64.7
Pixtral 12B
- 64.7 BlueLM-V-3B
- 64.6 MMAlaya2
- 64.4 Monkey-Chat
- 64.3
ShareGPT4V-13B
- 64.3
Eagle-X5-7B
- 64.2 InternVL-Chat-V1-5
- 63.8 llava-v1.6-vicuna-7b
- 63.8 Bunny-Llama-3-8B-V
- 63.4
LLaVA-v1.5-13B
- 63.4
ShareGPT4V-7B
- 63.4 LLaVA-v1.5-13B (QLoRA)
- 63.3 Aquila-VL-2B
- 63.2 LLaVA-Next-Qwen-32B
- 62.9 InternLM-XComposer2
- 62.6 Mantis-8B-Idefics2
- 61.9
Phi-3-vision-128k-instruct
- 61.4
Falcon2-11B-VLM
- 61.3 LLaVA-LLaMA-3-8B
- 61.2 LLaVA-InternLM2-20B (QLoRA)
- 61.2 llava-v1.6-mistral-7b
- 61.0 LLaVA-v1.5-7B (QLoRA)
- 60.9 TransCore-M
- 60.7 llama3-llava-next-8b
- 60.6 llava-v1.5-7b
- 60.6
LLaVA-NeXT-Interleave-7B
- 60.2
Phi-4-multimodal
- 59.6 Moondream1
- 59.1
OmniLMM-12B
- 58.6 Qwen-VL-Chat
- 58.5 Mantis-8B-siglip-llama3
- 58.1
DeepSeek-VL-1.3B-Chat
- 58.1 POINTS-Qwen2.5-7B
- 58.1 AKI-4B
- 58.0 LLaVA-InternLM2-7B (QLoRA)
- 57.8
PaliGemma-3B-mix-448
- 57.6 LLaVA-InternLM-7B (QLoRA)
- 57.3
LLaVA-NeXT-Interleave-7B-DPO
- 56.9
LLaVA-OneVision 0.5B (SI)
- 56.8 Vintern-3B-beta
- 56.4
MiniCPM-V
- 56.3 POINTS-Yi-1.5-9B
- 55.8
MiniCPM-V 2.0
- 55.0 Qwen-VL-Max
- 54.5 InternLM-XComposer
- 54.2 Mantis-8B-Fuyu
- 51.0 MMAlaya
- 51.0 Slime-13B
- 50.2 instructblip-vicuna-7b
- 50.2 ShareCaptioner
- 50.0 H2OVL-800M
- 49.5
Yi-VL-6B
- 49.0 Slime-7B
- 47.9 Slime-8B
- 43.7
InternVL2-Llama3-76B
- 43.4
InternVL2-26B
- 42.9
InternVL2-40B
- 42.6
InternVL2-8B
- 42.4
Yi-VL-34B
- 41.0 Emu2-Chat
- 38.9
InternVL3-8B
- 38.8
Qwen-VL-Plus
- 36.0 GPT-4.1 nano
- 30.5
InternVL3-14B
- 28.6
MolmoE-1B
- 27.6
Gemma 3 4B
- 26.3 GPT-4.1 mini
- 21.4
InternVL3-1B
- 19.1
InternVL3-2B
- 18.7 GPT-4.1
- 18.5
Gemini 1.5 Flash
- 17.7
Gemma 3 12B
- 15.9
Molmo-7B-D
- 15.1
Molmo-7B-O
- 13.5
Gemma 3 27B
- 12.3
Gemini 1.0 Pro
- 11.7
Janus-1.3B
- 9.00
Chameleon-7B
- 8.90 OpenFlamingo v2
- 3.70
Chameleon-30B
- 2.90
Kosmos-2
- 0.50 SmolVLM-Synthetic
- 0.30 MiniGPT-4-v1-7B
- 0.20
SmolVLM2-256M-Video-Instruct
- 0.10
SmolVLM-Instruct
- 0.10 SmolVLM-Instruct-DPO
- 0.0 VisualGLM
- 0.0 PandaGPT-13B
- 0.0
IDEFICS-9B-Instruct
- 0.0
IDEFICS-80B-Instruct
- 0.0 MiniGPT-4-v2
- 0.0 GPT-4v (1106, detail-low)
- 0.0
Idefics3-8B-Llama3
- 0.0
SmolVLM-256M-Instruct
- 0.0
SmolVLM-500M-Instruct
- 0.0
SmolVLM2-2.2B-Instruct
- 0.0
SmolVLM2-500M-Video-Instruct
- 1 Kimi-VL-A3B-Instruct ✓ 82%
- 2 MUG-U-7B ✓ 80.9%
- 3
Ovis1.6-Gemma2-27B ✓ 75.1%
- 4 Qwen2-VL-72B-Instruct ✓ 73.7%
- 5 Taiyi ✓ 73.5%
- 6
Ovis1.5-Llama3-8B ✓ 71.2%
- 7
Ovis1.6-Gemma2-9B ✓ 71.2%
- 8 H2OVL-2B ✓ 70.3%
- 9
Cambrian-1-34B ✓ 68.2%
- 10 Qwen-VL ✓ 67.5%
- 11 Monkey ✓ 67%
- 12
DeepSeek-VL-7B-Chat ✓ 67%
- 13
Ovis1.6-Llama3.2-3B ✓ 66.7%
- 14 CogVLM-17B-Chat ✓ 66.3%
- 15
LLaVA-v1.6-34B ✓ 66.3%
- 16
Cambrian-1-8B ✓ 66%
- 17 llava-v1.6-vicuna-13b ✓ 65.3%
- 18
mPLUG-Owl2 ✓ 65.2%
- 19
Eagle-X5-13B ✓ 65%
- 20 CogVLM2-19B-Chat ✓ 64.9%
- 21
LLaVA-OneVision 7B (SI) SI ✓ 64.7%
- 22
Pixtral 12B ✓ 64.7%
- 23 BlueLM-V-3B ✓ 64.7%
- 24 MMAlaya2 ✓ 64.6%
- 25 Monkey-Chat ✓ 64.4%
- 26
ShareGPT4V-13B ✓ 64.3%
- 27
Eagle-X5-7B ✓ 64.3%
- 28 InternVL-Chat-V1-5 ✓ 64.2%
- 29 llava-v1.6-vicuna-7b ✓ 63.8%
- 30 Bunny-Llama-3-8B-V ✓ 63.8%
- 31
LLaVA-v1.5-13B ✓ 63.4%
- 32
ShareGPT4V-7B ✓ 63.4%
- 33 LLaVA-v1.5-13B (QLoRA) QLoRA ✓ 63.4%
- 34 Aquila-VL-2B ✓ 63.3%
- 35 LLaVA-Next-Qwen-32B ✓ 63.2%
- 36 InternLM-XComposer2 ✓ 62.9%
- 37 Mantis-8B-Idefics2 ✓ 62.6%
- 38
Phi-3-vision-128k-instruct ✓ 61.9%
- 39
Falcon2-11B-VLM ✓ 61.4%
- 40 LLaVA-LLaMA-3-8B ✓ 61.3%
- 41 LLaVA-InternLM2-20B (QLoRA) QLoRA ✓ 61.2%
- 42 llava-v1.6-mistral-7b ✓ 61.2%
- 43 LLaVA-v1.5-7B (QLoRA) QLoRA ✓ 61%
- 44 TransCore-M ✓ 60.9%
- 45 llama3-llava-next-8b ✓ 60.7%
- 46 llava-v1.5-7b ✓ 60.6%
- 47
LLaVA-NeXT-Interleave-7B ✓ 60.6%
- 48
Phi-4-multimodal ✓ 60.2%
- 49 Moondream1 ✓ 59.6%
- 50
OmniLMM-12B ✓ 59.1%
- 51 Qwen-VL-Chat ✓ 58.6%
- 52 Mantis-8B-siglip-llama3 ✓ 58.5%
- 53
DeepSeek-VL-1.3B-Chat ✓ 58.1%
- 54 POINTS-Qwen2.5-7B ✓ 58.1%
- 55 AKI-4B ✓ 58.1%
- 56 LLaVA-InternLM2-7B (QLoRA) QLoRA ✓ 58%
- 57
PaliGemma-3B-mix-448 ✓ 57.8%
- 58 LLaVA-InternLM-7B (QLoRA) QLoRA ✓ 57.6%
- 59
LLaVA-NeXT-Interleave-7B-DPO ✓ 57.3%
- 60
LLaVA-OneVision 0.5B (SI) SI ✓ 56.9%
- 61 Vintern-3B-beta ✓ 56.8%
- 62
MiniCPM-V ✓ 56.4%
- 63 POINTS-Yi-1.5-9B ✓ 56.3%
- 64
MiniCPM-V 2.0 ✓ 55.8%
- 65 Qwen-VL-Max ✓ 55%
- 66 InternLM-XComposer ✓ 54.5%
- 67 Mantis-8B-Fuyu ✓ 54.2%
- 68 MMAlaya ✓ 51%
- 69 Slime-13B ✓ 51%
- 70 instructblip-vicuna-7b ✓ 50.2%
- 71 ShareCaptioner ✓ 50.2%
- 72 H2OVL-800M ✓ 50%
- 73
Yi-VL-6B ✓ 49.5%
- 74 Slime-7B ✓ 49%
- 75 Slime-8B ✓ 47.9%
- 76
InternVL2-Llama3-76B ✓ 43.7%
- 77
InternVL2-26B ✓ 43.4%
- 78
InternVL2-40B ✓ 42.9%
- 79
InternVL2-8B ✓ 42.6%
- 80
Yi-VL-34B ✓ 42.4%
- 81 Emu2-Chat ✓ 41%
- 82
InternVL3-8B ✓ 38.9%
- 83
Qwen-VL-Plus ✓ 38.8%
- 84 GPT-4.1 nano ✓ 36%
- 85
InternVL3-14B ✓ 30.5%
- 86
MolmoE-1B ✓ 28.6%
- 87
Gemma 3 4B ✓ 27.6%
- 88 GPT-4.1 mini ✓ 26.3%
- 89
InternVL3-1B ✓ 21.4%
- 90
InternVL3-2B ✓ 19.1%
- 91 GPT-4.1 ✓ 18.7%
- 92
Gemini 1.5 Flash Sep 2024 ✓ 18.5%
- 93
Gemma 3 12B ✓ 17.7%
- 94
Molmo-7B-D ✓ 15.9%
- 95
Molmo-7B-O ✓ 15.1%
- 96
Gemma 3 27B ✓ 13.5%
- 97
Gemini 1.0 Pro ✓ 12.3%
- 98
Janus-1.3B ✓ 11.7%
- 99
Chameleon-7B ✓ 9.00%
- 100 OpenFlamingo v2 8.90%
- 101
Chameleon-30B ✓ 3.70%
- 102
Kosmos-2 ✓ 2.90%
- 103 SmolVLM-Synthetic ✓ 0.50%
- 104 MiniGPT-4-v1-7B ✓ 0.30%
- 105
SmolVLM2-256M-Video-Instruct ✓ 0.20%
- 106
SmolVLM-Instruct ✓ 0.10%
- 107 SmolVLM-Instruct-DPO ✓ 0.10%
- 108 VisualGLM ✓ 0%
- 109 PandaGPT-13B ✓ 0%
- 110
IDEFICS-9B-Instruct ✓ 0%
- 111
IDEFICS-80B-Instruct ✓ 0%
- 112 MiniGPT-4-v2 ✓ 0%
- 113 GPT-4v (1106, detail-low) detail-low ✓ 0%
- 114
Idefics3-8B-Llama3 ✓ 0%
- 115
SmolVLM-256M-Instruct ✓ 0%
- 116
SmolVLM-500M-Instruct ✓ 0%
- 117
SmolVLM2-2.2B-Instruct ✓ 0%
- 118
SmolVLM2-500M-Video-Instruct ✓ 0%
Source: OpenCompass’s leaderboard
- 82.0 Kimi-VL-A3B-Instruct
- 75.1
Ovis1.6-Gemma2-27B
- 73.7 Qwen2-VL-72B-Instruct
- 71.2
Ovis1.5-Llama3-8B
- 71.2
Ovis1.6-Gemma2-9B
- 68.2
Cambrian-1-34B
- 67.0 Monkey
- 67.0
DeepSeek-VL-7B-Chat
- 66.7
Ovis1.6-Llama3.2-3B
- 66.3 CogVLM-17B-Chat
- 66.3
LLaVA-v1.6-34B
- 66.0
Cambrian-1-8B
- 65.3 llava-v1.6-vicuna-13b
- 65.2
mPLUG-Owl2
- 65.0
Eagle-X5-13B
- 64.9 CogVLM2-19B-Chat
- 64.7
LLaVA-OneVision 7B (SI)
- 64.7
Pixtral 12B
- 64.4 Monkey-Chat
- 64.3
ShareGPT4V-13B
- 64.3
Eagle-X5-7B
- 64.2 InternVL-Chat-V1-5
- 63.8 llava-v1.6-vicuna-7b
- 63.8 Bunny-Llama-3-8B-V
- 63.4
LLaVA-v1.5-13B
- 63.4
ShareGPT4V-7B
- 63.3 Aquila-VL-2B
- 61.9
Phi-3-vision-128k-instruct
- 61.4
Falcon2-11B-VLM
- 61.2 llava-v1.6-mistral-7b
- 60.7 llama3-llava-next-8b
- 60.6 llava-v1.5-7b
- 60.6
LLaVA-NeXT-Interleave-7B
- 60.2
Phi-4-multimodal
- 59.1
OmniLMM-12B
- 58.6 Qwen-VL-Chat
- 58.1
DeepSeek-VL-1.3B-Chat
- 57.8
PaliGemma-3B-mix-448
- 57.3
LLaVA-NeXT-Interleave-7B-DPO
- 56.9
LLaVA-OneVision 0.5B (SI)
- 56.4
MiniCPM-V
- 55.8
MiniCPM-V 2.0
- 55.0 Qwen-VL-Max
- 50.2 instructblip-vicuna-7b
- 49.5
Yi-VL-6B
- 43.7
InternVL2-Llama3-76B
- 43.4
InternVL2-26B
- 42.9
InternVL2-40B
- 42.6
InternVL2-8B
- 42.4
Yi-VL-34B
- 41.0 Emu2-Chat
- 38.9
InternVL3-8B
- 38.8
Qwen-VL-Plus
- 36.0 GPT-4.1 nano
- 30.5
InternVL3-14B
- 28.6
MolmoE-1B
- 27.6
Gemma 3 4B
- 26.3 GPT-4.1 mini
- 21.4
InternVL3-1B
- 19.1
InternVL3-2B
- 18.7 GPT-4.1
- 18.5
Gemini 1.5 Flash
- 17.7
Gemma 3 12B
- 15.9
Molmo-7B-D
- 15.1
Molmo-7B-O
- 13.5
Gemma 3 27B
- 12.3
Gemini 1.0 Pro
- 11.7
Janus-1.3B
- 9.00
Chameleon-7B
- 3.70
Chameleon-30B
- 2.90
Kosmos-2
- 0.20
SmolVLM2-256M-Video-Instruct
- 0.10
SmolVLM-Instruct
- 0.0
IDEFICS-9B-Instruct
- 0.0
IDEFICS-80B-Instruct
- 0.0
Idefics3-8B-Llama3
- 0.0
SmolVLM-256M-Instruct
- 0.0
SmolVLM-500M-Instruct
- 0.0
SmolVLM2-2.2B-Instruct
- 0.0
SmolVLM2-500M-Video-Instruct
- 1 Kimi-VL-A3B-Instruct ✓ 82%
- 2
Ovis1.6-Gemma2-27B ✓ 75.1%
- 3 Qwen2-VL-72B-Instruct ✓ 73.7%
- 4
Ovis1.5-Llama3-8B ✓ 71.2%
- 5
Ovis1.6-Gemma2-9B ✓ 71.2%
- 6
Cambrian-1-34B ✓ 68.2%
- 7 Monkey ✓ 67%
- 8
DeepSeek-VL-7B-Chat ✓ 67%
- 9
Ovis1.6-Llama3.2-3B ✓ 66.7%
- 10 CogVLM-17B-Chat ✓ 66.3%
- 11
LLaVA-v1.6-34B ✓ 66.3%
- 12
Cambrian-1-8B ✓ 66%
- 13 llava-v1.6-vicuna-13b ✓ 65.3%
- 14
mPLUG-Owl2 ✓ 65.2%
- 15
Eagle-X5-13B ✓ 65%
- 16 CogVLM2-19B-Chat ✓ 64.9%
- 17
LLaVA-OneVision 7B (SI) SI ✓ 64.7%
- 18
Pixtral 12B ✓ 64.7%
- 19 Monkey-Chat ✓ 64.4%
- 20
ShareGPT4V-13B ✓ 64.3%
- 21
Eagle-X5-7B ✓ 64.3%
- 22 InternVL-Chat-V1-5 ✓ 64.2%
- 23 llava-v1.6-vicuna-7b ✓ 63.8%
- 24 Bunny-Llama-3-8B-V ✓ 63.8%
- 25
LLaVA-v1.5-13B ✓ 63.4%
- 26
ShareGPT4V-7B ✓ 63.4%
- 27 Aquila-VL-2B ✓ 63.3%
- 28
Phi-3-vision-128k-instruct ✓ 61.9%
- 29
Falcon2-11B-VLM ✓ 61.4%
- 30 llava-v1.6-mistral-7b ✓ 61.2%
- 31 llama3-llava-next-8b ✓ 60.7%
- 32 llava-v1.5-7b ✓ 60.6%
- 33
LLaVA-NeXT-Interleave-7B ✓ 60.6%
- 34
Phi-4-multimodal ✓ 60.2%
- 35
OmniLMM-12B ✓ 59.1%
- 36 Qwen-VL-Chat ✓ 58.6%
- 37
DeepSeek-VL-1.3B-Chat ✓ 58.1%
- 38
PaliGemma-3B-mix-448 ✓ 57.8%
- 39
LLaVA-NeXT-Interleave-7B-DPO ✓ 57.3%
- 40
LLaVA-OneVision 0.5B (SI) SI ✓ 56.9%
- 41
MiniCPM-V ✓ 56.4%
- 42
MiniCPM-V 2.0 ✓ 55.8%
- 43 Qwen-VL-Max ✓ 55%
- 44 instructblip-vicuna-7b ✓ 50.2%
- 45
Yi-VL-6B ✓ 49.5%
- 46
InternVL2-Llama3-76B ✓ 43.7%
- 47
InternVL2-26B ✓ 43.4%
- 48
InternVL2-40B ✓ 42.9%
- 49
InternVL2-8B ✓ 42.6%
- 50
Yi-VL-34B ✓ 42.4%
- 51 Emu2-Chat ✓ 41%
- 52
InternVL3-8B ✓ 38.9%
- 53
Qwen-VL-Plus ✓ 38.8%
- 54 GPT-4.1 nano ✓ 36%
- 55
InternVL3-14B ✓ 30.5%
- 56
MolmoE-1B ✓ 28.6%
- 57
Gemma 3 4B ✓ 27.6%
- 58 GPT-4.1 mini ✓ 26.3%
- 59
InternVL3-1B ✓ 21.4%
- 60
InternVL3-2B ✓ 19.1%
- 61 GPT-4.1 ✓ 18.7%
- 62
Gemini 1.5 Flash Sep 2024 ✓ 18.5%
- 63
Gemma 3 12B ✓ 17.7%
- 64
Molmo-7B-D ✓ 15.9%
- 65
Molmo-7B-O ✓ 15.1%
- 66
Gemma 3 27B ✓ 13.5%
- 67
Gemini 1.0 Pro ✓ 12.3%
- 68
Janus-1.3B ✓ 11.7%
- 69
Chameleon-7B ✓ 9.00%
- 70
Chameleon-30B ✓ 3.70%
- 71
Kosmos-2 ✓ 2.90%
- 72
SmolVLM2-256M-Video-Instruct ✓ 0.20%
- 73
SmolVLM-Instruct ✓ 0.10%
- 74
IDEFICS-9B-Instruct ✓ 0%
- 75
IDEFICS-80B-Instruct ✓ 0%
- 76
Idefics3-8B-Llama3 ✓ 0%
- 77
SmolVLM-256M-Instruct ✓ 0%
- 78
SmolVLM-500M-Instruct ✓ 0%
- 79
SmolVLM2-2.2B-Instruct ✓ 0%
- 80
SmolVLM2-500M-Video-Instruct ✓ 0%
Source: OpenCompass’s leaderboard
| # | Entry | Score | Setting | Released | Run on | Link | Source |
|---|---|---|---|---|---|---|---|
| 1 | Kimi-VL-A3B-Instruct ✓ | 82% | VLMEvalKit | Apr 2025 | Apr 2025 | huggingface.co | Official leaderboard |
| 2 | MUG-U-7B ✓ closed weights, 7B | 80.9% | VLMEvalKit | Mar 2025 | github.com | Official leaderboard | |
| 3 | | 75.1% | VLMEvalKit | Nov 2024 | Nov 2024 | huggingface.co | Official leaderboard |
| 4 | Qwen2-VL-72B-Instruct ✓ | 73.7% | VLMEvalKit | Aug 2024 | Oct 2024 | huggingface.co | Official leaderboard |
| 5 | Taiyi ✓ closed weights | 73.5% | VLMEvalKit | Dec 2024 | taiyi.megvii.com | Official leaderboard | |
| 6 | | 71.2% | VLMEvalKit | Jul 2024 | Jul 2024 | huggingface.co | Official leaderboard |
| 7 | | 71.2% | VLMEvalKit | Sep 2024 | Sep 2024 | huggingface.co | Official leaderboard |
| 8 | H2OVL-2B ✓ open weights, 2.2B | 70.3% | VLMEvalKit | Oct 2024 | huggingface.co | Official leaderboard | |
| 9 | | 68.2% | VLMEvalKit | Jun 2024 | Jul 2024 | huggingface.co | Official leaderboard |
| 10 | Qwen-VL ✓ open weights, 9.6B | 67.5% | VLMEvalKit | Aug 2023 | github.com | Official leaderboard | |
| 11 | Monkey ✓ | 67% | VLMEvalKit | Nov 2023 | Jan 2024 | github.com | Official leaderboard |
| 12 | | 67% | VLMEvalKit | Mar 2024 | Mar 2024 | github.com | Official leaderboard |
| 13 | | 66.7% | VLMEvalKit | Oct 2024 | Oct 2024 | huggingface.co | Official leaderboard |
| 14 | CogVLM-17B-Chat ✓ | 66.3% | VLMEvalKit | Nov 2023 | Jan 2024 | huggingface.co | Official leaderboard |
| 15 | | 66.3% | VLMEvalKit | Jan 2024 | Mar 2024 | llava-vl.github.io | Official leaderboard |
| 16 | | 66% | VLMEvalKit | Jun 2024 | Jul 2024 | huggingface.co | Official leaderboard |
| 17 | llava-v1.6-vicuna-13b ✓ | 65.3% | VLMEvalKit | Jan 2024 | Mar 2024 | llava-vl.github.io | Official leaderboard |
| 18 | | 65.2% | VLMEvalKit | Nov 2023 | Nov 2023 | github.com | Official leaderboard |
| 19 | | 65% | VLMEvalKit | Aug 2024 | Sep 2024 | huggingface.co | Official leaderboard |
| 20 | CogVLM2-19B-Chat ✓ | 64.9% | VLMEvalKit | May 2024 | May 2024 | huggingface.co | Official leaderboard |
| 21 | | 64.7% | SI, VLMEvalKit | Aug 2024 | Sep 2024 | huggingface.co | Official leaderboard |
| 22 | | 64.7% | VLMEvalKit | Sep 2024 | Sep 2024 | huggingface.co | Official leaderboard |
| 23 | BlueLM-V-3B ✓ closed weights, 3B | 64.7% | VLMEvalKit | Oct 2024 | developers.vivo.com | Official leaderboard | |
| 24 | MMAlaya2 ✓ open weights, 26B | 64.6% | VLMEvalKit | Aug 2024 | huggingface.co | Official leaderboard | |
| 25 | Monkey-Chat ✓ | 64.4% | VLMEvalKit | Jan 2024 | Jan 2024 | github.com | Official leaderboard |
| 26 | | 64.3% | VLMEvalKit | Dec 2023 | Dec 2023 | sharegpt4v.github.io | Official leaderboard |
| 27 | | 64.3% | VLMEvalKit | Aug 2024 | Sep 2024 | huggingface.co | Official leaderboard |
| 28 | InternVL-Chat-V1-5 ✓ | 64.2% | VLMEvalKit | Apr 2024 | Apr 2024 | github.com | Official leaderboard |
| 29 | llava-v1.6-vicuna-7b ✓ | 63.8% | VLMEvalKit | Jan 2024 | Mar 2024 | llava-vl.github.io | Official leaderboard |
| 30 | Bunny-Llama-3-8B-V ✓ | 63.8% | VLMEvalKit | Apr 2024 | Jul 2024 | huggingface.co | Official leaderboard |
| 31 | | 63.4% | VLMEvalKit | Oct 2023 | Oct 2023 | arxiv.org | Official leaderboard |
| 32 | | 63.4% | VLMEvalKit | Nov 2023 | Dec 2023 | sharegpt4v.github.io | Official leaderboard |
| 33 | LLaVA-v1.5-13B (QLoRA) ✓ open weights, 13.4B | 63.4% | QLoRA, VLMEvalKit | Dec 2023 | github.com | Official leaderboard | |
| 34 | Aquila-VL-2B ✓ | 63.3% | VLMEvalKit | Oct 2024 | Oct 2024 | huggingface.co | Official leaderboard |
| 35 | LLaVA-Next-Qwen-32B ✓ open weights, 33B | 63.2% | VLMEvalKit | Aug 2024 | llava-vl.github.io | Official leaderboard | |
| 36 | InternLM-XComposer2 ✓ open weights, 7B | 62.9% | VLMEvalKit | Jan 2024 | github.com | Official leaderboard | |
| 37 | Mantis-8B-Idefics2 ✓ open weights, 8B | 62.6% | VLMEvalKit | Aug 2024 | huggingface.co | Official leaderboard | |
| 38 | | 61.9% | VLMEvalKit | May 2024 | May 2024 | huggingface.co | Official leaderboard |
| 39 | | 61.4% | VLMEvalKit | May 2024 | Oct 2024 | huggingface.co | Official leaderboard |
| 40 | LLaVA-LLaMA-3-8B ✓ open weights, 8B | 61.3% | VLMEvalKit | May 2024 | huggingface.co | Official leaderboard | |
| 41 | LLaVA-InternLM2-20B (QLoRA) ✓ open weights, 20.2B | 61.2% | QLoRA, VLMEvalKit | Jan 2024 | github.com | Official leaderboard | |
| 42 | llava-v1.6-mistral-7b ✓ | 61.2% | VLMEvalKit | Jan 2024 | Mar 2024 | llava-vl.github.io | Official leaderboard |
| 43 | LLaVA-v1.5-7B (QLoRA) ✓ open weights, 7.2B | 61% | QLoRA, VLMEvalKit | Dec 2023 | github.com | Official leaderboard | |
| 44 | TransCore-M ✓ open weights, 13.4B | 60.9% | VLMEvalKit | Apr 2024 | github.com | Official leaderboard | |
| 45 | llama3-llava-next-8b ✓ | 60.7% | VLMEvalKit | Apr 2024 | Jul 2024 | llava-vl.github.io | Official leaderboard |
| 46 | llava-v1.5-7b ✓ | 60.6% | VLMEvalKit | Oct 2023 | Oct 2023 | arxiv.org | Official leaderboard |
| 47 | | 60.6% | VLMEvalKit | Jun 2024 | Aug 2024 | llava-vl.github.io | Official leaderboard |
| 48 | | 60.2% | VLMEvalKit | Feb 2025 | Mar 2025 | huggingface.co | Official leaderboard |
| 49 | Moondream1 ✓ open weights, 1.9B | 59.6% | VLMEvalKit | Sep 2024 | huggingface.co | Official leaderboard | |
| 50 | | 59.1% | VLMEvalKit | Jan 2024 | Feb 2024 | huggingface.co | Official leaderboard |
| 51 | Qwen-VL-Chat ✓ | 58.6% | VLMEvalKit | Aug 2023 | Aug 2023 | github.com | Official leaderboard |
| 52 | Mantis-8B-siglip-llama3 ✓ open weights, 8B | 58.5% | VLMEvalKit | Aug 2024 | huggingface.co | Official leaderboard | |
| 53 | | 58.1% | VLMEvalKit | Mar 2024 | Mar 2024 | github.com | Official leaderboard |
| 54 | POINTS-Qwen2.5-7B ✓ open weights, 8.3B | 58.1% | VLMEvalKit | Oct 2024 | huggingface.co | Official leaderboard | |
| 55 | AKI-4B ✓ open weights, 4.3B | 58.1% | VLMEvalKit | Mar 2025 | huggingface.co | Official leaderboard | |
| 56 | LLaVA-InternLM2-7B (QLoRA) ✓ open weights, 8.1B | 58% | QLoRA, VLMEvalKit | Jan 2024 | github.com | Official leaderboard | |
| 57 | | 57.8% | VLMEvalKit | May 2024 | May 2024 | github.com | Official leaderboard |
| 58 | LLaVA-InternLM-7B (QLoRA) ✓ open weights, 7.6B | 57.6% | QLoRA, VLMEvalKit | Dec 2023 | github.com | Official leaderboard | |
| 59 | | 57.3% | VLMEvalKit | Jun 2024 | Jul 2024 | llava-vl.github.io | Official leaderboard |
| 60 | | 56.9% | SI, VLMEvalKit | Aug 2024 | Sep 2024 | huggingface.co | Official leaderboard |
| 61 | Vintern-3B-beta ✓ open weights, 3.7B | 56.8% | VLMEvalKit | Oct 2024 | huggingface.co | Official leaderboard | |
| 62 | | 56.4% | VLMEvalKit | Jan 2024 | Feb 2024 | huggingface.co | Official leaderboard |
| 63 | POINTS-Yi-1.5-9B ✓ open weights, 9.5B | 56.3% | VLMEvalKit | Oct 2024 | huggingface.co | Official leaderboard | |
| 64 | | 55.8% | VLMEvalKit | Apr 2024 | Apr 2024 | huggingface.co | Official leaderboard |
| 65 | Qwen-VL-Max ✓ | 55% | VLMEvalKit | Jan 2024 | Feb 2024 | huggingface.co | Official leaderboard |
| 66 | InternLM-XComposer ✓ open weights, 8B | 54.5% | VLMEvalKit | Sep 2023 | github.com | Official leaderboard | |
| 67 | Mantis-8B-Fuyu ✓ open weights, 8B | 54.2% | VLMEvalKit | Sep 2024 | huggingface.co | Official leaderboard | |
| 68 | MMAlaya ✓ open weights, 7.8B | 51% | VLMEvalKit | Jan 2024 | huggingface.co | Official leaderboard | |
| 69 | Slime-13B ✓ open weights, 13B | 51% | VLMEvalKit | Sep 2024 | huggingface.co | Official leaderboard | |
| 70 | instructblip-vicuna-7b ✓ | 50.2% | VLMEvalKit | May 2023 | May 2023 | arxiv.org | Official leaderboard |
| 71 | ShareCaptioner ✓ open weights, 8B | 50.2% | VLMEvalKit | Jan 2024 | huggingface.co | Official leaderboard | |
| 72 | H2OVL-800M ✓ open weights, 0.8B | 50% | VLMEvalKit | Oct 2024 | huggingface.co | Official leaderboard | |
| 73 | | 49.5% | VLMEvalKit | Dec 2023 | Jan 2024 | huggingface.co | Official leaderboard |
| 74 | Slime-7B ✓ open weights, 7B | 49% | VLMEvalKit | Sep 2024 | huggingface.co | Official leaderboard | |
| 75 | Slime-8B ✓ open weights, 8B | 47.9% | VLMEvalKit | Sep 2024 | huggingface.co | Official leaderboard | |
| 76 | | 43.7% | VLMEvalKit | Jul 2024 | Jul 2024 | huggingface.co | Official leaderboard |
| 77 | | 43.4% | VLMEvalKit | Jul 2024 | Jul 2024 | huggingface.co | Official leaderboard |
| 78 | | 42.9% | VLMEvalKit | Jul 2024 | Jul 2024 | huggingface.co | Official leaderboard |
| 79 | | 42.6% | VLMEvalKit | Jul 2024 | Jul 2024 | huggingface.co | Official leaderboard |
| 80 | | 42.4% | VLMEvalKit | Dec 2023 | Jan 2024 | huggingface.co | Official leaderboard |
| 81 | Emu2-Chat ✓ | 41% | VLMEvalKit | Dec 2023 | Jan 2024 | github.com | Official leaderboard |
| 82 | | 38.9% | VLMEvalKit | Apr 2025 | Apr 2025 | huggingface.co | Official leaderboard |
| 83 | | 38.8% | VLMEvalKit | Jan 2024 | Jan 2024 | github.com | Official leaderboard |
| 84 | GPT-4.1 nano ✓ | 36% | VLMEvalKit | Apr 2025 | Apr 2025 | openai.com | Official leaderboard |
| 85 | | 30.5% | VLMEvalKit | Apr 2025 | Apr 2025 | huggingface.co | Official leaderboard |
| 86 | | 28.6% | VLMEvalKit | Sep 2024 | Oct 2024 | huggingface.co | Official leaderboard |
| 87 | | 27.6% | VLMEvalKit | Mar 2025 | Mar 2025 | huggingface.co | Official leaderboard |
| 88 | GPT-4.1 mini ✓ | 26.3% | VLMEvalKit | Apr 2025 | Apr 2025 | openai.com | Official leaderboard |
| 89 | | 21.4% | VLMEvalKit | Apr 2025 | Apr 2025 | huggingface.co | Official leaderboard |
| 90 | | 19.1% | VLMEvalKit | Apr 2025 | Apr 2025 | huggingface.co | Official leaderboard |
| 91 | GPT-4.1 ✓ | 18.7% | VLMEvalKit | Apr 2025 | Apr 2025 | openai.com | Official leaderboard |
| 92 | | 18.5% | VLMEvalKit | Aug 2024 | Sep 2024 | deepmind.google | Official leaderboard |
| 93 | | 17.7% | VLMEvalKit | Mar 2025 | Mar 2025 | huggingface.co | Official leaderboard |
| 94 | | 15.9% | VLMEvalKit | Sep 2024 | Oct 2024 | huggingface.co | Official leaderboard |
| 95 | | 15.1% | VLMEvalKit | Sep 2024 | Oct 2024 | huggingface.co | Official leaderboard |
| 96 | | 13.5% | VLMEvalKit | Mar 2025 | Mar 2025 | huggingface.co | Official leaderboard |
| 97 | | 12.3% | VLMEvalKit | Dec 2023 | Dec 2023 | deepmind.google | Official leaderboard |
| 98 | | 11.7% | VLMEvalKit | Oct 2024 | Oct 2024 | huggingface.co | Official leaderboard |
| 99 | | 9.00% | VLMEvalKit | Jun 2024 | Jul 2024 | huggingface.co | Official leaderboard |
| 100 | OpenFlamingo v2 open weights, 9B | 8.90% | VLMEvalKit | Jun 2023 | github.com | Official leaderboard | |
| 101 | | 3.70% | VLMEvalKit | Jun 2024 | Jul 2024 | huggingface.co | Official leaderboard |
| 102 | | 2.90% | VLMEvalKit | Jun 2023 | Oct 2024 | huggingface.co | Official leaderboard |
| 103 | SmolVLM-Synthetic ✓ open weights, 2.3B | 0.50% | VLMEvalKit | Jan 2025 | huggingface.co | Official leaderboard | |
| 104 | MiniGPT-4-v1-7B ✓ open weights, 8B | 0.30% | VLMEvalKit | Apr 2023 | arxiv.org | Official leaderboard | |
| 105 | | 0.20% | VLMEvalKit | Feb 2025 | Mar 2025 | huggingface.co | Official leaderboard |
| 106 | | 0.10% | VLMEvalKit | Nov 2024 | Jan 2025 | huggingface.co | Official leaderboard |
| 107 | SmolVLM-Instruct-DPO ✓ open weights, 2.3B | 0.10% | VLMEvalKit | Jan 2025 | huggingface.co | Official leaderboard | |
| 108 | VisualGLM ✓ open weights, 8B | 0% | VLMEvalKit | May 2023 | github.com | Official leaderboard | |
| 109 | PandaGPT-13B ✓ open weights, 14B | 0% | VLMEvalKit | May 2023 | arxiv.org | Official leaderboard | |
| 110 | | 0% | VLMEvalKit | Jul 2023 | Jul 2023 | huggingface.co | Official leaderboard |
| 111 | | 0% | VLMEvalKit | Jul 2023 | Jul 2023 | huggingface.co | Official leaderboard |
| 112 | MiniGPT-4-v2 ✓ open weights, 8B | 0% | VLMEvalKit | Oct 2023 | minigpt-v2.github.io | Official leaderboard | |
| 113 | GPT-4v (1106, detail-low) ✓ closed weights | 0% | detail-low, VLMEvalKit | Dec 2023 | openai.com | Official leaderboard | |
| 114 | | 0% | VLMEvalKit | Aug 2024 | Aug 2024 | huggingface.co | Official leaderboard |
| 115 | | 0% | VLMEvalKit | Jan 2025 | Jan 2025 | huggingface.co | Official leaderboard |
| 116 | | 0% | VLMEvalKit | Jan 2025 | Jan 2025 | huggingface.co | Official leaderboard |
| 117 | | 0% | VLMEvalKit | Feb 2025 | Mar 2025 | huggingface.co | Official leaderboard |
| 118 | | 0% | VLMEvalKit | Feb 2025 | Mar 2025 | huggingface.co | Official leaderboard |
| # | Model | Score | Setting | Released | Run on | Link | Source |
|---|---|---|---|---|---|---|---|
| 1 | Kimi-VL-A3B-Instruct ✓ | 82% | VLMEvalKit | Apr 2025 | Apr 2025 | huggingface.co | Official leaderboard |
| 2 | | 75.1% | VLMEvalKit | Nov 2024 | Nov 2024 | huggingface.co | Official leaderboard |
| 3 | Qwen2-VL-72B-Instruct ✓ | 73.7% | VLMEvalKit | Aug 2024 | Oct 2024 | huggingface.co | Official leaderboard |
| 4 | | 71.2% | VLMEvalKit | Jul 2024 | Jul 2024 | huggingface.co | Official leaderboard |
| 5 | | 71.2% | VLMEvalKit | Sep 2024 | Sep 2024 | huggingface.co | Official leaderboard |
| 6 | | 68.2% | VLMEvalKit | Jun 2024 | Jul 2024 | huggingface.co | Official leaderboard |
| 7 | Monkey ✓ | 67% | VLMEvalKit | Nov 2023 | Jan 2024 | github.com | Official leaderboard |
| 8 | | 67% | VLMEvalKit | Mar 2024 | Mar 2024 | github.com | Official leaderboard |
| 9 | | 66.7% | VLMEvalKit | Oct 2024 | Oct 2024 | huggingface.co | Official leaderboard |
| 10 | CogVLM-17B-Chat ✓ | 66.3% | VLMEvalKit | Nov 2023 | Jan 2024 | huggingface.co | Official leaderboard |
| 11 | | 66.3% | VLMEvalKit | Jan 2024 | Mar 2024 | llava-vl.github.io | Official leaderboard |
| 12 | | 66% | VLMEvalKit | Jun 2024 | Jul 2024 | huggingface.co | Official leaderboard |
| 13 | llava-v1.6-vicuna-13b ✓ | 65.3% | VLMEvalKit | Jan 2024 | Mar 2024 | llava-vl.github.io | Official leaderboard |
| 14 | | 65.2% | VLMEvalKit | Nov 2023 | Nov 2023 | github.com | Official leaderboard |
| 15 | | 65% | VLMEvalKit | Aug 2024 | Sep 2024 | huggingface.co | Official leaderboard |
| 16 | CogVLM2-19B-Chat ✓ | 64.9% | VLMEvalKit | May 2024 | May 2024 | huggingface.co | Official leaderboard |
| 17 | | 64.7% | SI, VLMEvalKit | Aug 2024 | Sep 2024 | huggingface.co | Official leaderboard |
| 18 | | 64.7% | VLMEvalKit | Sep 2024 | Sep 2024 | huggingface.co | Official leaderboard |
| 19 | Monkey-Chat ✓ | 64.4% | VLMEvalKit | Jan 2024 | Jan 2024 | github.com | Official leaderboard |
| 20 | | 64.3% | VLMEvalKit | Dec 2023 | Dec 2023 | sharegpt4v.github.io | Official leaderboard |
| 21 | | 64.3% | VLMEvalKit | Aug 2024 | Sep 2024 | huggingface.co | Official leaderboard |
| 22 | InternVL-Chat-V1-5 ✓ | 64.2% | VLMEvalKit | Apr 2024 | Apr 2024 | github.com | Official leaderboard |
| 23 | llava-v1.6-vicuna-7b ✓ | 63.8% | VLMEvalKit | Jan 2024 | Mar 2024 | llava-vl.github.io | Official leaderboard |
| 24 | Bunny-Llama-3-8B-V ✓ | 63.8% | VLMEvalKit | Apr 2024 | Jul 2024 | huggingface.co | Official leaderboard |
| 25 | | 63.4% | VLMEvalKit | Oct 2023 | Oct 2023 | arxiv.org | Official leaderboard |
| 26 | | 63.4% | VLMEvalKit | Nov 2023 | Dec 2023 | sharegpt4v.github.io | Official leaderboard |
| 27 | Aquila-VL-2B ✓ | 63.3% | VLMEvalKit | Oct 2024 | Oct 2024 | huggingface.co | Official leaderboard |
| 28 | | 61.9% | VLMEvalKit | May 2024 | May 2024 | huggingface.co | Official leaderboard |
| 29 | | 61.4% | VLMEvalKit | May 2024 | Oct 2024 | huggingface.co | Official leaderboard |
| 30 | llava-v1.6-mistral-7b ✓ | 61.2% | VLMEvalKit | Jan 2024 | Mar 2024 | llava-vl.github.io | Official leaderboard |
| 31 | llama3-llava-next-8b ✓ | 60.7% | VLMEvalKit | Apr 2024 | Jul 2024 | llava-vl.github.io | Official leaderboard |
| 32 | llava-v1.5-7b ✓ | 60.6% | VLMEvalKit | Oct 2023 | Oct 2023 | arxiv.org | Official leaderboard |
| 33 | | 60.6% | VLMEvalKit | Jun 2024 | Aug 2024 | llava-vl.github.io | Official leaderboard |
| 34 | | 60.2% | VLMEvalKit | Feb 2025 | Mar 2025 | huggingface.co | Official leaderboard |
| 35 | | 59.1% | VLMEvalKit | Jan 2024 | Feb 2024 | huggingface.co | Official leaderboard |
| 36 | Qwen-VL-Chat ✓ | 58.6% | VLMEvalKit | Aug 2023 | Aug 2023 | github.com | Official leaderboard |
| 37 | | 58.1% | VLMEvalKit | Mar 2024 | Mar 2024 | github.com | Official leaderboard |
| 38 | | 57.8% | VLMEvalKit | May 2024 | May 2024 | github.com | Official leaderboard |
| 39 | | 57.3% | VLMEvalKit | Jun 2024 | Jul 2024 | llava-vl.github.io | Official leaderboard |
| 40 | | 56.9% | SI, VLMEvalKit | Aug 2024 | Sep 2024 | huggingface.co | Official leaderboard |
| 41 | | 56.4% | VLMEvalKit | Jan 2024 | Feb 2024 | huggingface.co | Official leaderboard |
| 42 | | 55.8% | VLMEvalKit | Apr 2024 | Apr 2024 | huggingface.co | Official leaderboard |
| 43 | Qwen-VL-Max ✓ | 55% | VLMEvalKit | Jan 2024 | Feb 2024 | huggingface.co | Official leaderboard |
| 44 | instructblip-vicuna-7b ✓ | 50.2% | VLMEvalKit | May 2023 | May 2023 | arxiv.org | Official leaderboard |
| 45 | | 49.5% | VLMEvalKit | Dec 2023 | Jan 2024 | huggingface.co | Official leaderboard |
| 46 | | 43.7% | VLMEvalKit | Jul 2024 | Jul 2024 | huggingface.co | Official leaderboard |
| 47 | | 43.4% | VLMEvalKit | Jul 2024 | Jul 2024 | huggingface.co | Official leaderboard |
| 48 | | 42.9% | VLMEvalKit | Jul 2024 | Jul 2024 | huggingface.co | Official leaderboard |
| 49 | | 42.6% | VLMEvalKit | Jul 2024 | Jul 2024 | huggingface.co | Official leaderboard |
| 50 | | 42.4% | VLMEvalKit | Dec 2023 | Jan 2024 | huggingface.co | Official leaderboard |
| 51 | Emu2-Chat ✓ | 41% | VLMEvalKit | Dec 2023 | Jan 2024 | github.com | Official leaderboard |
| 52 | | 38.9% | VLMEvalKit | Apr 2025 | Apr 2025 | huggingface.co | Official leaderboard |
| 53 | | 38.8% | VLMEvalKit | Jan 2024 | Jan 2024 | github.com | Official leaderboard |
| 54 | GPT-4.1 nano ✓ | 36% | VLMEvalKit | Apr 2025 | Apr 2025 | openai.com | Official leaderboard |
| 55 | | 30.5% | VLMEvalKit | Apr 2025 | Apr 2025 | huggingface.co | Official leaderboard |
| 56 | | 28.6% | VLMEvalKit | Sep 2024 | Oct 2024 | huggingface.co | Official leaderboard |
| 57 | | 27.6% | VLMEvalKit | Mar 2025 | Mar 2025 | huggingface.co | Official leaderboard |
| 58 | GPT-4.1 mini ✓ | 26.3% | VLMEvalKit | Apr 2025 | Apr 2025 | openai.com | Official leaderboard |
| 59 | | 21.4% | VLMEvalKit | Apr 2025 | Apr 2025 | huggingface.co | Official leaderboard |
| 60 | | 19.1% | VLMEvalKit | Apr 2025 | Apr 2025 | huggingface.co | Official leaderboard |
| 61 | GPT-4.1 ✓ | 18.7% | VLMEvalKit | Apr 2025 | Apr 2025 | openai.com | Official leaderboard |
| 62 | | 18.5% | VLMEvalKit | Aug 2024 | Sep 2024 | deepmind.google | Official leaderboard |
| 63 | | 17.7% | VLMEvalKit | Mar 2025 | Mar 2025 | huggingface.co | Official leaderboard |
| 64 | | 15.9% | VLMEvalKit | Sep 2024 | Oct 2024 | huggingface.co | Official leaderboard |
| 65 | | 15.1% | VLMEvalKit | Sep 2024 | Oct 2024 | huggingface.co | Official leaderboard |
| 66 | | 13.5% | VLMEvalKit | Mar 2025 | Mar 2025 | huggingface.co | Official leaderboard |
| 67 | | 12.3% | VLMEvalKit | Dec 2023 | Dec 2023 | deepmind.google | Official leaderboard |
| 68 | | 11.7% | VLMEvalKit | Oct 2024 | Oct 2024 | huggingface.co | Official leaderboard |
| 69 | | 9.00% | VLMEvalKit | Jun 2024 | Jul 2024 | huggingface.co | Official leaderboard |
| 70 | | 3.70% | VLMEvalKit | Jun 2024 | Jul 2024 | huggingface.co | Official leaderboard |
| 71 | | 2.90% | VLMEvalKit | Jun 2023 | Oct 2024 | huggingface.co | Official leaderboard |
| 72 | | 0.20% | VLMEvalKit | Feb 2025 | Mar 2025 | huggingface.co | Official leaderboard |
| 73 | | 0.10% | VLMEvalKit | Nov 2024 | Jan 2025 | huggingface.co | Official leaderboard |
| 74 | | 0% | VLMEvalKit | Jul 2023 | Jul 2023 | huggingface.co | Official leaderboard |
| 75 | | 0% | VLMEvalKit | Jul 2023 | Jul 2023 | huggingface.co | Official leaderboard |
| 76 | | 0% | VLMEvalKit | Aug 2024 | Aug 2024 | huggingface.co | Official leaderboard |
| 77 | | 0% | VLMEvalKit | Jan 2025 | Jan 2025 | huggingface.co | Official leaderboard |
| 78 | | 0% | VLMEvalKit | Jan 2025 | Jan 2025 | huggingface.co | Official leaderboard |
| 79 | | 0% | VLMEvalKit | Feb 2025 | Mar 2025 | huggingface.co | Official leaderboard |
| 80 | | 0% | VLMEvalKit | Feb 2025 | Mar 2025 | huggingface.co | Official leaderboard |
Rankings by part
We read 32 parts from OpenCompass’s leaderboard, each ranking models on one part of the test, such as Arts & Photography, Biographies & Memoirs and Business & Money. Kimi-VL-A3B-Instruct leads 27 of them, and 2 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
Arts & Photography
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 87.5%
- 2
Ovis1.6-Gemma2-27B 80.2%
- 3
Ovis1.6-Gemma2-9B ↑2 78.1%
- 4
Ovis1.5-Llama3-8B 77.1%
- 5
Phi-3-vision-128k-instruct ↑23 75%
- 6 Qwen2-VL-72B-Instruct ↓3 74%
- 7
Ovis1.6-Llama3.2-3B ↑2 70.8%
- 8 Aquila-VL-2B ↑19 70.8%
- 9
LLaVA-v1.6-34B ↑2 69.8%
- 10 InternVL-Chat-V1-5 ↑12 69.8%
Biographies & Memoirs
80 models we track have a score here.
- 1
Ovis1.6-Gemma2-27B ↑1 81.2%
- 2 Kimi-VL-A3B-Instruct ↓1 80.2%
- 3
Ovis1.5-Llama3-8B ↑1 77.1%
- 4
LLaVA-v1.5-13B ↑21 75%
- 5
LLaVA-v1.6-34B ↑6 75%
- 6
Ovis1.6-Gemma2-9B ↓1 75%
- 7 Qwen2-VL-72B-Instruct ↓4 74%
- 8
Pixtral 12B ↑10 74%
- 9
mPLUG-Owl2 ↑5 72.9%
- 10 llava-v1.6-vicuna-7b ↑13 72.9%
Business & Money
80 models we track have a score here.
- 1 Qwen2-VL-72B-Instruct ↑2 84.4%
- 2
Ovis1.6-Gemma2-27B 84.4%
- 3 Monkey ↑4 83.3%
- 4 Monkey-Chat ↑15 83.3%
- 5 Kimi-VL-A3B-Instruct ↓4 83.3%
- 6 CogVLM2-19B-Chat ↑10 82.3%
- 7
Ovis1.5-Llama3-8B ↓3 82.3%
- 8
Ovis1.6-Gemma2-9B ↓3 82.3%
- 9
mPLUG-Owl2 ↑5 81.2%
- 10
Cambrian-1-8B ↑2 81.2%
Calendars
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 81.2%
- 2
Ovis1.5-Llama3-8B ↑2 74%
- 3
Ovis1.6-Gemma2-9B ↑2 72.9%
- 4
Ovis1.6-Gemma2-27B ↓2 71.9%
- 5
Cambrian-1-34B ↑1 69.8%
- 6
DeepSeek-VL-7B-Chat ↑2 68.8%
- 7
Eagle-X5-7B ↑14 68.8%
- 8 Qwen2-VL-72B-Instruct ↓5 68.8%
- 9
Ovis1.6-Llama3.2-3B 68.8%
- 10
LLaVA-v1.5-13B ↑15 67.7%
Children's Books
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 85.4%
- 2 Qwen2-VL-72B-Instruct ↑1 81.2%
- 3
Ovis1.6-Gemma2-27B ↓1 80.2%
- 4
Ovis1.5-Llama3-8B 78.1%
- 5
Cambrian-1-34B ↑1 75%
- 6
Ovis1.6-Gemma2-9B ↓1 75%
- 7
Cambrian-1-8B ↑5 74%
- 8
Ovis1.6-Llama3.2-3B ↑1 74%
- 9
Eagle-X5-13B ↑6 72.9%
- 10 CogVLM-17B-Chat 71.9%
Christian Books & Bibles
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 90.6%
- 2
Ovis1.6-Gemma2-27B 85.4%
- 3
ShareGPT4V-13B ↑17 79.2%
- 4 llava-v1.6-vicuna-13b ↑9 79.2%
- 5
Ovis1.5-Llama3-8B ↓1 79.2%
- 6
Ovis1.6-Gemma2-9B ↓1 79.2%
- 7
LLaVA-v1.5-13B ↑18 78.1%
- 8
LLaVA-v1.6-34B ↑3 78.1%
- 9 Qwen2-VL-72B-Instruct ↓6 78.1%
- 10 Monkey ↓3 77.1%
Comics & Graphic Novels
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 82.3%
- 2
Ovis1.6-Gemma2-27B 77.1%
- 3
InternVL3-8B ↑49 77.1%
- 4
DeepSeek-VL-7B-Chat ↑4 76%
- 5 Qwen2-VL-72B-Instruct ↓2 75%
- 6 llava-v1.6-vicuna-13b ↑7 72.9%
- 7
Cambrian-1-34B ↓1 72.9%
- 8
Ovis1.5-Llama3-8B ↓4 72.9%
- 9
InternVL3-14B ↑46 71.9%
- 10 InternVL-Chat-V1-5 ↑12 70.8%
Computers & Technology
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 92.7%
- 2
Ovis1.6-Gemma2-27B 78.1%
- 3
DeepSeek-VL-7B-Chat ↑5 77.1%
- 4 Qwen2-VL-72B-Instruct ↓1 77.1%
- 5
LLaVA-v1.6-34B ↑6 76%
- 6 Monkey ↑1 75%
- 7
Cambrian-1-34B ↓1 74%
- 8
Cambrian-1-8B ↑4 74%
- 9
Ovis1.6-Gemma2-9B ↓4 74%
- 10
LLaVA-v1.5-13B ↑15 72.9%
Cookbooks, Food & Wine
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 84.4%
- 2 Qwen2-VL-72B-Instruct ↑1 75%
- 3
Ovis1.6-Gemma2-27B ↓1 75%
- 4
Ovis1.6-Gemma2-9B ↑1 72.9%
- 5
mPLUG-Owl2 ↑9 71.9%
- 6
Ovis1.5-Llama3-8B ↓2 71.9%
- 7
Pixtral 12B ↑11 71.9%
- 8 CogVLM-17B-Chat ↑2 70.8%
- 9
LLaVA-v1.6-34B ↑2 70.8%
- 10 Bunny-Llama-3-8B-V ↑14 70.8%
Crafts, Hobbies & Home
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 83.3%
- 2 Qwen2-VL-72B-Instruct ↑1 79.2%
- 3
Cambrian-1-34B ↑3 75%
- 4
Ovis1.6-Gemma2-27B ↓2 75%
- 5
Ovis1.5-Llama3-8B ↓1 74%
- 6
Ovis1.6-Gemma2-9B ↓1 71.9%
- 7 CogVLM-17B-Chat ↑3 70.8%
- 8 Monkey ↓1 69.8%
- 9 Monkey-Chat ↑10 69.8%
- 10 CogVLM2-19B-Chat ↑6 69.8%
Education & Teaching
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 77.1%
- 2
Ovis1.6-Gemma2-27B 70.8%
- 3 Qwen2-VL-72B-Instruct 69.8%
- 4
Ovis1.6-Gemma2-9B ↑1 67.7%
- 5
Cambrian-1-34B ↑1 64.6%
- 6 Monkey ↑1 63.5%
- 7
Cambrian-1-8B ↑5 63.5%
- 8
LLaVA-v1.6-34B ↑3 62.5%
- 9
Ovis1.5-Llama3-8B ↓5 62.5%
- 10
LLaVA-OneVision 7B (SI) ↑7 62.5%
Engineering & Transportation
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 71.9%
- 2
Ovis1.6-Gemma2-27B 64.6%
- 3
Ovis1.6-Gemma2-9B ↑2 63.5%
- 4 Qwen2-VL-72B-Instruct ↓1 60.4%
- 5 Monkey ↑2 57.3%
- 6 Qwen-VL-Chat ↑30 56.2%
- 7
DeepSeek-VL-7B-Chat ↑1 56.2%
- 8
Ovis1.5-Llama3-8B ↓4 54.2%
- 9 CogVLM-17B-Chat ↑1 53.1%
- 10 Monkey-Chat ↑9 53.1%
Gay & Lesbian
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 74%
- 2 Qwen2-VL-72B-Instruct ↑1 71.9%
- 3
Ovis1.5-Llama3-8B ↑1 67.7%
- 4
Ovis1.6-Gemma2-27B ↓2 66.7%
- 5
Ovis1.6-Gemma2-9B 63.5%
- 6 Bunny-Llama-3-8B-V ↑18 62.5%
- 7
mPLUG-Owl2 ↑7 61.5%
- 8
Cambrian-1-8B ↑4 61.5%
- 9 CogVLM-17B-Chat ↑1 60.4%
- 10 Monkey ↓3 60.4%
Health, Fitness & Dieting
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 80.2%
- 2 Qwen2-VL-72B-Instruct ↑1 66.7%
- 3
Ovis1.5-Llama3-8B ↑1 65.6%
- 4
Ovis1.6-Gemma2-9B ↑1 63.5%
- 5
Ovis1.6-Gemma2-27B ↓3 63.5%
- 6
Ovis1.6-Llama3.2-3B ↑3 62.5%
- 7 CogVLM-17B-Chat ↑3 60.4%
- 8 CogVLM2-19B-Chat ↑8 60.4%
- 9
mPLUG-Owl2 ↑5 59.4%
- 10 Monkey-Chat ↑9 59.4%
History
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 93.8%
- 2
DeepSeek-VL-7B-Chat ↑6 78.1%
- 3
Ovis1.6-Gemma2-27B ↓1 78.1%
- 4
Ovis1.5-Llama3-8B 76%
- 5 Qwen2-VL-72B-Instruct ↓2 75%
- 6 Monkey ↑1 74%
- 7
Ovis1.6-Llama3.2-3B ↑2 72.9%
- 8 llama3-llava-next-8b ↑23 71.9%
- 9 CogVLM-17B-Chat ↑1 70.8%
- 10
ShareGPT4V-7B ↑16 70.8%
Humor & Entertainment
80 models we track have a score here.
- 1
Ovis1.6-Gemma2-27B ↑1 80.2%
- 2 Kimi-VL-A3B-Instruct ↓1 80.2%
- 3 Qwen2-VL-72B-Instruct 78.1%
- 4
Ovis1.6-Gemma2-9B ↑1 78.1%
- 5
Ovis1.5-Llama3-8B ↓1 76%
- 6
LLaVA-NeXT-Interleave-7B ↑27 72.9%
- 7 CogVLM-17B-Chat ↑3 71.9%
- 8 Monkey-Chat ↑11 69.8%
- 9 llava-v1.6-vicuna-13b ↑4 69.8%
- 10
Ovis1.6-Llama3.2-3B ↓1 69.8%
Law
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 76%
- 2 Qwen2-VL-72B-Instruct ↑1 70.8%
- 3
Ovis1.5-Llama3-8B ↑1 68.8%
- 4
Ovis1.6-Gemma2-27B ↓2 68.8%
- 5
Ovis1.6-Gemma2-9B 65.6%
- 6
LLaVA-v1.6-34B ↑5 64.6%
- 7
DeepSeek-VL-7B-Chat ↑1 64.6%
- 8
mPLUG-Owl2 ↑6 63.5%
- 9 Monkey ↓2 63.5%
- 10 llava-v1.6-vicuna-7b ↑13 63.5%
Literature & Fiction
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 87.5%
- 2
Ovis1.6-Gemma2-27B 80.2%
- 3
Ovis1.6-Gemma2-9B ↑2 76%
- 4
Cambrian-1-8B ↑8 75%
- 5 Qwen2-VL-72B-Instruct ↓2 72.9%
- 6
Cambrian-1-34B 71.9%
- 7
ShareGPT4V-7B ↑19 69.8%
- 8
ShareGPT4V-13B ↑12 69.8%
- 9
Eagle-X5-13B ↑6 69.8%
- 10
LLaVA-v1.5-13B ↑15 68.8%
Medical Books
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 79.2%
- 2 Qwen2-VL-72B-Instruct ↑1 69.8%
- 3
Ovis1.6-Gemma2-27B ↓1 69.8%
- 4
Ovis1.6-Gemma2-9B ↑1 67.7%
- 5 Monkey ↑2 65.6%
- 6 CogVLM-17B-Chat ↑4 63.5%
- 7
Ovis1.5-Llama3-8B ↓3 63.5%
- 8 Aquila-VL-2B ↑19 62.5%
- 9
ShareGPT4V-13B ↑11 61.5%
- 10 llava-v1.6-vicuna-13b ↑3 61.5%
Mystery, Thriller & Suspense
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 81.2%
- 2
DeepSeek-VL-7B-Chat ↑6 76%
- 3 llava-v1.6-vicuna-13b ↑10 71.9%
- 4
Ovis1.6-Gemma2-9B ↑1 71.9%
- 5
Cambrian-1-34B ↑1 69.8%
- 6
LLaVA-v1.6-34B ↑5 68.8%
- 7 Bunny-Llama-3-8B-V ↑17 68.8%
- 8
Cambrian-1-8B ↑4 68.8%
- 9
Ovis1.6-Gemma2-27B ↓7 68.8%
- 10 InternVL-Chat-V1-5 ↑12 67.7%
Parenting & Relationships
80 models we track have a score here.
- 1 Qwen2-VL-72B-Instruct ↑2 77.1%
- 2
Ovis1.6-Gemma2-27B 77.1%
- 3 Kimi-VL-A3B-Instruct ↓2 70.8%
- 4
Ovis1.6-Gemma2-9B ↑1 69.8%
- 5 Monkey ↑2 67.7%
- 6 Monkey-Chat ↑13 67.7%
- 7
Ovis1.6-Llama3.2-3B ↑2 67.7%
- 8
Ovis1.5-Llama3-8B ↓4 66.7%
- 9
PaliGemma-3B-mix-448 ↑29 65.6%
- 10
Cambrian-1-34B ↓4 65.6%
Politics & Social Sciences
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 76%
- 2
Ovis1.6-Gemma2-27B 71.9%
- 3 Qwen2-VL-72B-Instruct 70.8%
- 4
DeepSeek-VL-7B-Chat ↑4 68.8%
- 5
Ovis1.5-Llama3-8B ↓1 67.7%
- 6
Ovis1.6-Gemma2-9B ↓1 67.7%
- 7
LLaVA-v1.5-13B ↑18 63.5%
- 8 CogVLM-17B-Chat ↑2 63.5%
- 9 Monkey ↓2 63.5%
- 10
ShareGPT4V-13B ↑10 62.5%
Reference
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 72.9%
- 2
Ovis1.6-Gemma2-27B 66.7%
- 3 Qwen2-VL-72B-Instruct 62.5%
- 4
Ovis1.6-Gemma2-9B ↑1 61.5%
- 5
Ovis1.5-Llama3-8B ↓1 58.3%
- 6 Monkey ↑1 57.3%
- 7 Monkey-Chat ↑12 56.2%
- 8
DeepSeek-VL-7B-Chat 55.2%
- 9
Cambrian-1-34B ↓3 55.2%
- 10
Eagle-X5-13B ↑5 55.2%
Religion & Spirituality
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 78.1%
- 2 Qwen2-VL-72B-Instruct ↑1 76%
- 3 Monkey ↑4 68.8%
- 4
Ovis1.6-Gemma2-27B ↓2 68.8%
- 5
ShareGPT4V-7B ↑21 67.7%
- 6
MiniCPM-V 2.0 ↑36 67.7%
- 7
Ovis1.5-Llama3-8B ↓3 66.7%
- 8
Pixtral 12B ↑10 66.7%
- 9 Monkey-Chat ↑10 65.6%
- 10 llava-v1.6-vicuna-13b ↑3 65.6%
Romance
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 86.5%
- 2 Qwen2-VL-72B-Instruct ↑1 82.3%
- 3
Ovis1.6-Gemma2-9B ↑2 81.2%
- 4
Ovis1.6-Gemma2-27B ↓2 81.2%
- 5
Ovis1.5-Llama3-8B ↓1 80.2%
- 6
Cambrian-1-34B 78.1%
- 7
Eagle-X5-13B ↑8 78.1%
- 8
mPLUG-Owl2 ↑6 77.1%
- 9 CogVLM-17B-Chat ↑1 77.1%
- 10
Ovis1.6-Llama3.2-3B ↓1 77.1%
Science & Math
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 89.6%
- 2
Ovis1.6-Gemma2-27B 83.3%
- 3
Cambrian-1-34B ↑3 79.2%
- 4
Cambrian-1-8B ↑8 78.1%
- 5
Ovis1.5-Llama3-8B ↓1 78.1%
- 6
Eagle-X5-13B ↑9 78.1%
- 7
Ovis1.6-Gemma2-9B ↓2 78.1%
- 8 CogVLM-17B-Chat ↑2 76%
- 9
LLaVA-v1.6-34B ↑2 76%
- 10
DeepSeek-VL-7B-Chat ↓2 75%
Science Fiction & Fantasy
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 86.5%
- 2 Qwen2-VL-72B-Instruct ↑1 76%
- 3
Ovis1.6-Gemma2-27B ↓1 75%
- 4 Monkey ↑3 68.8%
- 5
Ovis1.5-Llama3-8B ↓1 68.8%
- 6 CogVLM-17B-Chat ↑4 66.7%
- 7
Cambrian-1-34B ↓1 66.7%
- 8
Ovis1.6-Gemma2-9B ↓3 66.7%
- 9 Monkey-Chat ↑10 65.6%
- 10
LLaVA-v1.5-13B ↑15 64.6%
Self-Help
80 models we track have a score here.
- 1
Ovis1.6-Gemma2-27B ↑1 77.1%
- 2 Kimi-VL-A3B-Instruct ↓1 77.1%
- 3
DeepSeek-VL-7B-Chat ↑5 76%
- 4
Ovis1.6-Llama3.2-3B ↑5 75%
- 5
Ovis1.5-Llama3-8B ↓1 74%
- 6
Pixtral 12B ↑12 74%
- 7 Qwen2-VL-72B-Instruct ↓4 72.9%
- 8
Ovis1.6-Gemma2-9B ↓3 72.9%
- 9 InternVL-Chat-V1-5 ↑13 69.8%
- 10
Phi-3-vision-128k-instruct ↑18 69.8%
Sports & Outdoors
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 87.5%
- 2 Qwen2-VL-72B-Instruct ↑1 74%
- 3
Ovis1.6-Gemma2-27B ↓1 74%
- 4
Cambrian-1-34B ↑2 69.8%
- 5
Cambrian-1-8B ↑7 68.8%
- 6 Monkey ↑1 66.7%
- 7
LLaVA-v1.6-34B ↑4 66.7%
- 8
Ovis1.5-Llama3-8B ↓4 66.7%
- 9 llava-v1.6-vicuna-13b ↑4 65.6%
- 10
Eagle-X5-13B ↑5 65.6%
Teen & Young Adult
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 79.2%
- 2
Ovis1.6-Gemma2-27B 75%
- 3 Qwen2-VL-72B-Instruct 71.9%
- 4 CogVLM-17B-Chat ↑6 70.8%
- 5
Ovis1.5-Llama3-8B ↓1 70.8%
- 6 CogVLM2-19B-Chat ↑10 66.7%
- 7
Cambrian-1-34B ↓1 65.6%
- 8
LLaVA-OneVision 7B (SI) ↑9 64.6%
- 9
Ovis1.6-Gemma2-9B ↓4 64.6%
- 10
Ovis1.6-Llama3.2-3B ↓1 64.6%
Test Preparation
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 88.5%
- 2 Qwen2-VL-72B-Instruct ↑1 77.1%
- 3
Ovis1.6-Gemma2-27B ↓1 77.1%
- 4
Ovis1.5-Llama3-8B 75%
- 5
DeepSeek-VL-7B-Chat ↑3 74%
- 6
LLaVA-v1.6-34B ↑5 72.9%
- 7
Cambrian-1-34B ↓1 72.9%
- 8
Pixtral 12B ↑10 72.9%
- 9
LLaVA-OneVision 7B (SI) ↑8 71.9%
- 10
Ovis1.6-Gemma2-9B ↓5 70.8%
Travel
80 models we track have a score here.
- 1 Kimi-VL-A3B-Instruct 80.2%
- 2
LLaVA-v1.6-34B ↑9 76%
- 3
Cambrian-1-34B ↑3 76%
- 4
Ovis1.5-Llama3-8B 76%
- 5 Qwen2-VL-72B-Instruct ↓2 76%
- 6
Ovis1.6-Gemma2-27B ↓4 76%
- 7
mPLUG-Owl2 ↑7 75%
- 8 CogVLM-17B-Chat ↑2 75%
- 9 llava-v1.6-vicuna-13b ↑4 75%
- 10
Cambrian-1-8B ↑2 75%
11 models have held number one on OpenCompass’s leaderboard since 2023
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: OpenCompass’s leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: OpenCompass’s leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Vision
- MMMU In use 11.5K college-level multimodal questions across 30 subjects. Multiple-choice and open accuracy.
Qwen 3.6 Plus 86%
- MMMU-Pro In use Harder MMMU: college-level multimodal questions with more options and vision-only inputs to remove text shortcuts. GPT-5.6 Sol 84.6%
- CharXiv In use Chart understanding on real arXiv figures. Reasoning questions require synthesizing information from complex charts. Claude Fable 5 93.5%
- MathVista In use Math reasoning in visual contexts (charts, geometry, puzzles) from 31 datasets. Answer accuracy.
Seed2.1 Pro 90.7%
- MATH-Vision (MathVision) In use 3,040 competition math problems with visual context. Answer accuracy.
Qwen 3.8 Max 97.7%
- MMStar In use Models answer 1,500 human-curated image questions that cannot be solved without the image, across 6 core capabilities. Scored by accuracy.
Qwen 3.8 Max 85.9%
Data audit
9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 19 months ago, in April 2025, for GPT-4.1. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
- Pass Independent runs Scores from the official leaderboard (118 entries).
- Pass Models scored 80 models we track, plus 38 other entries on its leaderboard.
- Warning Newest score The newest came 19 months ago, in April 2025, for GPT-4.1.
- Note Frontier models None of the top 10 has a score. The strongest tested is GPT-4.1 (#138).
- Pass Headroom The best is 82%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on OpenCompass: 80 rows for models we track and 38 other entries at the last read.
- Pass Last read November 20th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.