MCS-Bench
Models read seal inscriptions, identify calligraphy, and answer questions about Chinese paintings and artifacts. The overall score averages task scores.
The authors compare 37 models on 23 multiple-choice and 22 open-ended tasks. Institutions must apply for the full dataset.
- Made by
- South China University of Technology and INTSIG
- Released
- February 2025
- Kind
- General visual questions
- Level
- Expert
- Scored
- Checked automatically
- Answers
- Multiple choice
- Tasks
- Kept private
- Measures
- Capability
- Language
- Chinese
- Interaction
- Single turn
- Items from
- Written by experts, Real work and use
- Refresh
- One fixed set
- Run by
- A host
- In views
- Vision
- Strongest model tested
-
Gemini 2.0 Flash Feb 2025 #147 on the capability index
- Last model added
- March 2025
- Scores refreshed
- November 20th, 2026
InternVL2_5-78B leads with 49.3 among models released by March 2025 InternVL2_5-78B leads with 49.3 among models released by March 2025
Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. 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 March 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
- 49.3
InternVL2_5-78B
- 46.5
InternVL2.5-38B
- 46.2
Gemini 2.0 Flash
- 45.9
InternVL2.5-26B
- 44.7 GPT-4o
- 44.6 Qwen2-VL-72B-Instruct
- 44.5 Claude 3.5 Sonnet
- 43.9
Gemini 1.5 Pro
- 43.6
InternVL2.5-8B
- 43.3
InternVL2.5-4B
- 42.7
InternVL2-26B
- 42.5
QVQ-72B-Preview
- 41.8 GLM-4V-9B
- 41.0
InternVL2-8B
- 40.6 Qwen2-VL-7B-Instruct
- 39.1 MiniCPM-V-2_6
- 38.8
InternVL2.5-1B
- 38.8
InternVL2.5-2B
- 35.6
Qwen2-VL-2B-Instruct
- 34.8
InternVL2-4B
- 34.2
DeepSeek-VL2-Tiny
- 33.5 Qwen-VL-Chat
- 32.5
Ovis1.6-Gemma2-9B
- 32.5
MiniCPM-V 2.0
- 32.4
MiniCPM-Llama3-V 2.5
- 30.5
DeepSeek-VL2
- 30.4
MiniCPM-V
- 30.2
Ovis1.5-Gemma2-9B
- 26.3
DeepSeek-VL2-Small
- 24.5 llava-v1.6-vicuna-7b
- 24.1
LLaVA-v1.5-13B
- 23.6 llava-v1.6-vicuna-13b
- 23.4 llava-v1.6-mistral-7b
- 23.3 llava-v1.5-7b
- 21.2
Molmo-7B-D
- 21.0 Chinese-LLaVA-CLLaMA2
- 20.2
Molmo-7B-O
- 1
InternVL2_5-78B 49.3
- 2
InternVL2.5-38B 46.5
- 3
Gemini 2.0 Flash Feb 2025 46.2
- 4
InternVL2.5-26B 45.9
- 5 GPT-4o Mar 2025 44.7
- 6 Qwen2-VL-72B-Instruct 44.6
- 7 Claude 3.5 Sonnet 44.5
- 8
Gemini 1.5 Pro Sep 2024 43.9
- 9
InternVL2.5-8B 43.6
- 10
InternVL2.5-4B 43.3
- 11
InternVL2-26B 42.7
- 12
QVQ-72B-Preview 42.5
- 13 GLM-4V-9B 41.8
- 14
InternVL2-8B 41.0
- 15 Qwen2-VL-7B-Instruct 40.6
- 16 MiniCPM-V-2_6 39.1
- 17
InternVL2.5-1B 38.8
- 18
InternVL2.5-2B 38.8
- 19
Qwen2-VL-2B-Instruct 35.6
- 20
InternVL2-4B 34.8
- 21
DeepSeek-VL2-Tiny 34.2
- 22 Qwen-VL-Chat 33.5
- 23
Ovis1.6-Gemma2-9B 32.5
- 24
MiniCPM-V 2.0 32.5
- 25
MiniCPM-Llama3-V 2.5 32.4
- 26
DeepSeek-VL2 30.5
- 27
MiniCPM-V 30.4
- 28
Ovis1.5-Gemma2-9B 30.2
- 29
DeepSeek-VL2-Small 26.3
- 30 llava-v1.6-vicuna-7b 24.5
- 31
LLaVA-v1.5-13B 24.1
- 32 llava-v1.6-vicuna-13b 23.6
- 33 llava-v1.6-mistral-7b 23.4
- 34 llava-v1.5-7b 23.3
- 35
Molmo-7B-D 21.2
- 36 Chinese-LLaVA-CLLaMA2 21.0
- 37
Molmo-7B-O 20.2
Source: The MCS-Bench paper
- 49.3
InternVL2_5-78B
- 46.5
InternVL2.5-38B
- 46.2
Gemini 2.0 Flash
- 45.9
InternVL2.5-26B
- 44.7 GPT-4o
- 44.6 Qwen2-VL-72B-Instruct
- 44.5 Claude 3.5 Sonnet
- 43.9
Gemini 1.5 Pro
- 43.6
InternVL2.5-8B
- 43.3
InternVL2.5-4B
- 42.7
InternVL2-26B
- 42.5
QVQ-72B-Preview
- 41.8 GLM-4V-9B
- 41.0
InternVL2-8B
- 40.6 Qwen2-VL-7B-Instruct
- 39.1 MiniCPM-V-2_6
- 38.8
InternVL2.5-1B
- 38.8
InternVL2.5-2B
- 35.6
Qwen2-VL-2B-Instruct
- 34.8
InternVL2-4B
- 34.2
DeepSeek-VL2-Tiny
- 33.5 Qwen-VL-Chat
- 32.5
Ovis1.6-Gemma2-9B
- 32.5
MiniCPM-V 2.0
- 32.4
MiniCPM-Llama3-V 2.5
- 30.5
DeepSeek-VL2
- 30.4
MiniCPM-V
- 30.2
Ovis1.5-Gemma2-9B
- 26.3
DeepSeek-VL2-Small
- 24.5 llava-v1.6-vicuna-7b
- 24.1
LLaVA-v1.5-13B
- 23.6 llava-v1.6-vicuna-13b
- 23.4 llava-v1.6-mistral-7b
- 23.3 llava-v1.5-7b
- 21.2
Molmo-7B-D
- 20.2
Molmo-7B-O
- 1
InternVL2_5-78B 49.3
- 2
InternVL2.5-38B 46.5
- 3
Gemini 2.0 Flash Feb 2025 46.2
- 4
InternVL2.5-26B 45.9
- 5 GPT-4o Mar 2025 44.7
- 6 Qwen2-VL-72B-Instruct 44.6
- 7 Claude 3.5 Sonnet 44.5
- 8
Gemini 1.5 Pro Sep 2024 43.9
- 9
InternVL2.5-8B 43.6
- 10
InternVL2.5-4B 43.3
- 11
InternVL2-26B 42.7
- 12
QVQ-72B-Preview 42.5
- 13 GLM-4V-9B 41.8
- 14
InternVL2-8B 41.0
- 15 Qwen2-VL-7B-Instruct 40.6
- 16 MiniCPM-V-2_6 39.1
- 17
InternVL2.5-1B 38.8
- 18
InternVL2.5-2B 38.8
- 19
Qwen2-VL-2B-Instruct 35.6
- 20
InternVL2-4B 34.8
- 21
DeepSeek-VL2-Tiny 34.2
- 22 Qwen-VL-Chat 33.5
- 23
Ovis1.6-Gemma2-9B 32.5
- 24
MiniCPM-V 2.0 32.5
- 25
MiniCPM-Llama3-V 2.5 32.4
- 26
DeepSeek-VL2 30.5
- 27
MiniCPM-V 30.4
- 28
Ovis1.5-Gemma2-9B 30.2
- 29
DeepSeek-VL2-Small 26.3
- 30 llava-v1.6-vicuna-7b 24.5
- 31
LLaVA-v1.5-13B 24.1
- 32 llava-v1.6-vicuna-13b 23.6
- 33 llava-v1.6-mistral-7b 23.4
- 34 llava-v1.5-7b 23.3
- 35
Molmo-7B-D 21.2
- 36
Molmo-7B-O 20.2
Source: The MCS-Bench paper
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 49.3 | Dec 2024 | Official leaderboard | |
| 2 | | 46.5 | Dec 2024 | Official leaderboard | |
| 3 | | 46.2 | Feb 2025 | Official leaderboard | |
| 4 | | 45.9 | Dec 2024 | Official leaderboard | |
| 5 | GPT-4o Mar 2025 | 44.7 | Mar 2025 | Official leaderboard | |
| 6 | Qwen2-VL-72B-Instruct | 44.6 | Aug 2024 | Official leaderboard | |
| 7 | Claude 3.5 Sonnet | 44.5 | Jun 2024 | Official leaderboard | |
| 8 | | 43.9 | Sep 2024 | Official leaderboard | |
| 9 | | 43.6 | Dec 2024 | Official leaderboard | |
| 10 | | 43.3 | Dec 2024 | Official leaderboard | |
| 11 | | 42.7 | Jul 2024 | Official leaderboard | |
| 12 | | 42.5 | Dec 2024 | Official leaderboard | |
| 13 | GLM-4V-9B | 41.8 | Jun 2024 | Official leaderboard | |
| 14 | | 41.0 | Jul 2024 | Official leaderboard | |
| 15 | Qwen2-VL-7B-Instruct | 40.6 | Aug 2024 | Official leaderboard | |
| 16 | MiniCPM-V-2_6 | 39.1 | Aug 2024 | Official leaderboard | |
| 17 | | 38.8 | Dec 2024 | Official leaderboard | |
| 18 | | 38.8 | Dec 2024 | Official leaderboard | |
| 19 | | 35.6 | Aug 2024 | Official leaderboard | |
| 20 | | 34.8 | Jul 2024 | Official leaderboard | |
| 21 | | 34.2 | Dec 2024 | Official leaderboard | |
| 22 | Qwen-VL-Chat | 33.5 | Aug 2023 | Official leaderboard | |
| 23 | | 32.5 | Sep 2024 | Official leaderboard | |
| 24 | | 32.5 | Apr 2024 | Official leaderboard | |
| 25 | | 32.4 | May 2024 | Official leaderboard | |
| 26 | | 30.5 | Dec 2024 | Official leaderboard | |
| 27 | | 30.4 | Jan 2024 | Official leaderboard | |
| 28 | | 30.2 | Jul 2024 | Official leaderboard | |
| 29 | | 26.3 | Dec 2024 | Official leaderboard | |
| 30 | llava-v1.6-vicuna-7b | 24.5 | Jan 2024 | Official leaderboard | |
| 31 | | 24.1 | Oct 2023 | Official leaderboard | |
| 32 | llava-v1.6-vicuna-13b | 23.6 | Jan 2024 | Official leaderboard | |
| 33 | llava-v1.6-mistral-7b | 23.4 | Jan 2024 | Official leaderboard | |
| 34 | llava-v1.5-7b | 23.3 | Oct 2023 | Official leaderboard | |
| 35 | | 21.2 | Sep 2024 | Official leaderboard | |
| 36 | Chinese-LLaVA-CLLaMA2 open weights | 21.0 | Official leaderboard | ||
| 37 | | 20.2 | Sep 2024 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 49.3 | Dec 2024 | Official leaderboard | |
| 2 | | 46.5 | Dec 2024 | Official leaderboard | |
| 3 | | 46.2 | Feb 2025 | Official leaderboard | |
| 4 | | 45.9 | Dec 2024 | Official leaderboard | |
| 5 | GPT-4o Mar 2025 | 44.7 | Mar 2025 | Official leaderboard | |
| 6 | Qwen2-VL-72B-Instruct | 44.6 | Aug 2024 | Official leaderboard | |
| 7 | Claude 3.5 Sonnet | 44.5 | Jun 2024 | Official leaderboard | |
| 8 | | 43.9 | Sep 2024 | Official leaderboard | |
| 9 | | 43.6 | Dec 2024 | Official leaderboard | |
| 10 | | 43.3 | Dec 2024 | Official leaderboard | |
| 11 | | 42.7 | Jul 2024 | Official leaderboard | |
| 12 | | 42.5 | Dec 2024 | Official leaderboard | |
| 13 | GLM-4V-9B | 41.8 | Jun 2024 | Official leaderboard | |
| 14 | | 41.0 | Jul 2024 | Official leaderboard | |
| 15 | Qwen2-VL-7B-Instruct | 40.6 | Aug 2024 | Official leaderboard | |
| 16 | MiniCPM-V-2_6 | 39.1 | Aug 2024 | Official leaderboard | |
| 17 | | 38.8 | Dec 2024 | Official leaderboard | |
| 18 | | 38.8 | Dec 2024 | Official leaderboard | |
| 19 | | 35.6 | Aug 2024 | Official leaderboard | |
| 20 | | 34.8 | Jul 2024 | Official leaderboard | |
| 21 | | 34.2 | Dec 2024 | Official leaderboard | |
| 22 | Qwen-VL-Chat | 33.5 | Aug 2023 | Official leaderboard | |
| 23 | | 32.5 | Sep 2024 | Official leaderboard | |
| 24 | | 32.5 | Apr 2024 | Official leaderboard | |
| 25 | | 32.4 | May 2024 | Official leaderboard | |
| 26 | | 30.5 | Dec 2024 | Official leaderboard | |
| 27 | | 30.4 | Jan 2024 | Official leaderboard | |
| 28 | | 30.2 | Jul 2024 | Official leaderboard | |
| 29 | | 26.3 | Dec 2024 | Official leaderboard | |
| 30 | llava-v1.6-vicuna-7b | 24.5 | Jan 2024 | Official leaderboard | |
| 31 | | 24.1 | Oct 2023 | Official leaderboard | |
| 32 | llava-v1.6-vicuna-13b | 23.6 | Jan 2024 | Official leaderboard | |
| 33 | llava-v1.6-mistral-7b | 23.4 | Jan 2024 | Official leaderboard | |
| 34 | llava-v1.5-7b | 23.3 | Oct 2023 | Official leaderboard | |
| 35 | | 21.2 | Sep 2024 | Official leaderboard | |
| 36 | | 20.2 | Sep 2024 | Official leaderboard |
Rankings by part
We read 54 parts from the benchmark’s paper, each ranking models on one part of the test, such as Ancient Chinese Text, Calligraphy and Painting. InternVL2_5-78B leads 20 of them, and 12 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
Ancient Chinese Text
36 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑2 51.4
- 2
Gemini 1.5 Pro Sep 2024 ↑6 50.6
- 3
InternVL2_5-78B ↓2 42.9
- 4 Qwen2-VL-72B-Instruct ↑2 42.8
- 5
InternVL2.5-38B ↓3 39.5
- 6 Claude 3.5 Sonnet ↑1 38.5
- 7 Qwen2-VL-7B-Instruct ↑8 37.8
- 8
InternVL2.5-8B ↑1 36.9
- 9
InternVL2.5-26B ↓5 36.8
- 10 GPT-4o Mar 2025 ↓5 35.5
Calligraphy
36 models we track have a score here.
- 1
InternVL2_5-78B 51.6
- 2
InternVL2.5-26B ↑2 48.5
- 3
InternVL2.5-8B ↑6 47.6
- 4
InternVL2.5-4B ↑6 47.3
- 5
InternVL2.5-38B ↓3 47.1
- 6
QVQ-72B-Preview ↑6 45.4
- 7
InternVL2.5-1B ↑10 45.4
- 8
InternVL2-8B ↑6 45.3
- 9 GPT-4o Mar 2025 ↓4 44.8
- 10 Qwen2-VL-72B-Instruct ↓4 43.8
Painting
36 models we track have a score here.
- 1
InternVL2_5-78B 62.7
- 2
InternVL2.5-38B 61.3
- 3
InternVL2.5-26B ↑1 59.5
- 4
Gemini 2.0 Flash Feb 2025 ↓1 59.0
- 5
InternVL2.5-4B ↑5 58.9
- 6
Gemini 1.5 Pro Sep 2024 ↑2 58.9
- 7 Qwen2-VL-72B-Instruct ↓1 58.8
- 8 GPT-4o Mar 2025 ↓3 58.0
- 9 Claude 3.5 Sonnet ↓2 57.8
- 10
InternVL2.5-8B ↓1 57.1
Oracle Bone Script
36 models we track have a score here.
- 1
InternVL2_5-78B 27.5
- 2
InternVL2.5-38B 26.8
- 3
InternVL2.5-26B ↑1 26.1
- 4
MiniCPM-Llama3-V 2.5 ↑21 25.8
- 5 Qwen2-VL-7B-Instruct ↑10 25.7
- 6 GPT-4o Mar 2025 ↓1 25.5
- 7
InternVL2.5-4B ↑3 25.4
- 8 GLM-4V-9B ↑5 25.3
- 9
MiniCPM-V 2.0 ↑15 25.2
- 10
InternVL2.5-8B ↓1 25.2
Seal
36 models we track have a score here.
- 1
InternVL2_5-78B 29.0
- 2
InternVL2.5-38B 26.4
- 3
InternVL2-26B ↑8 25.4
- 4
InternVL2.5-26B 23.8
- 5 GLM-4V-9B ↑8 23.3
- 6
QVQ-72B-Preview ↑6 22.8
- 7
InternVL2.5-4B ↑3 22.2
- 8
Gemini 2.0 Flash Feb 2025 ↓5 21.9
- 9 Claude 3.5 Sonnet ↓2 21.4
- 10 GPT-4o Mar 2025 ↓5 21.4
Cultural Relic
36 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑2 69.2
- 2
InternVL2_5-78B ↓1 68.8
- 3 GPT-4o Mar 2025 ↑2 68.4
- 4 Claude 3.5 Sonnet ↑3 68.2
- 5 Qwen2-VL-72B-Instruct ↑1 65.6
- 6 GLM-4V-9B ↑7 65.4
- 7
InternVL2.5-26B ↓3 65.3
- 8
InternVL2.5-38B ↓6 64.0
- 9 MiniCPM-V-2_6 ↑7 63.9
- 10
QVQ-72B-Preview ↑2 63.3
Illustration
36 models we track have a score here.
- 1
InternVL2_5-78B 62.7
- 2
InternVL2.5-26B ↑2 61.4
- 3
Gemini 2.0 Flash Feb 2025 60.8
- 4
InternVL2.5-38B ↓2 60.7
- 5 Claude 3.5 Sonnet ↑2 60.3
- 6 GPT-4o Mar 2025 ↓1 59.5
- 7 Qwen2-VL-72B-Instruct ↓1 58.9
- 8
InternVL2-26B ↑3 58.5
- 9
InternVL2.5-8B 57.0
- 10
QVQ-72B-Preview ↑2 56.6
MCQ
36 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑2 65.1
- 2
InternVL2_5-78B ↓1 64.5
- 3 GPT-4o Mar 2025 ↑2 61.8
- 4 Claude 3.5 Sonnet ↑3 60.8
- 5
InternVL2.5-38B ↓3 60.6
- 6
Gemini 1.5 Pro Sep 2024 ↑2 60.5
- 7 Qwen2-VL-72B-Instruct ↓1 60.0
- 8
InternVL2.5-26B ↓4 59.7
- 9
QVQ-72B-Preview ↑3 58.8
- 10 GLM-4V-9B ↑3 57.3
QA
36 models we track have a score here.
- 1
InternVL2_5-78B 40.4
- 2
InternVL2.5-38B 38.6
- 3
InternVL2.5-26B ↑1 37.9
- 4
InternVL2.5-8B ↑5 36.7
- 5
InternVL2.5-4B ↑5 36.5
- 6 Qwen2-VL-72B-Instruct 34.9
- 7
InternVL2.5-1B ↑10 34.7
- 8 Claude 3.5 Sonnet ↓1 34.7
- 9
InternVL2-8B ↑5 34.4
- 10
InternVL2-26B ↑1 34.2
T1
36 models we track have a score here.
- 1
InternVL2_5-78B 75.2
- 2
Gemini 2.0 Flash Feb 2025 ↑1 72.2
- 3
InternVL2.5-38B ↓1 68.0
- 4
InternVL2.5-26B 63.8
- 5
Gemini 1.5 Pro Sep 2024 ↑3 61.8
- 6
InternVL2.5-8B ↑3 60.6
- 7 Qwen2-VL-72B-Instruct ↓1 59.5
- 8
InternVL2-8B ↑6 59.4
- 9
InternVL2-26B ↑2 56.7
- 10
InternVL2.5-4B 56.2
T2
36 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑2 10.4
- 2
Gemini 1.5 Pro Sep 2024 ↑6 10.1
- 3 Qwen2-VL-72B-Instruct ↑3 2.29
- 4 Claude 3.5 Sonnet ↑3 2.08
- 5
InternVL2_5-78B ↓4 2.05
- 6 GPT-4o Mar 2025 ↓1 1.91
- 7
InternVL2.5-26B ↓3 1.84
- 8 Qwen2-VL-7B-Instruct ↑7 1.51
- 9
InternVL2.5-4B ↑1 1.45
- 10 GLM-4V-9B ↑3 1.36
T3
36 models we track have a score here.
- 1
Gemini 1.5 Pro Sep 2024 ↑7 57.3
- 2
Gemini 2.0 Flash Feb 2025 ↑1 47.9
- 3 Claude 3.5 Sonnet ↑4 24.5
- 4
QVQ-72B-Preview ↑8 18.3
- 5 Qwen2-VL-72B-Instruct ↑1 12.8
- 6 GPT-4o Mar 2025 ↓1 12.6
- 7
InternVL2.5-26B ↓3 12.0
- 8 Qwen2-VL-7B-Instruct ↑7 9.73
- 9
InternVL2.5-38B ↓7 8.43
- 10
InternVL2_5-78B ↓9 8.42
T4
36 models we track have a score here.
- 1
Gemini 1.5 Pro Sep 2024 ↑7 36.9
- 2
Gemini 2.0 Flash Feb 2025 ↑1 32.2
- 3 Qwen2-VL-72B-Instruct ↑3 19.9
- 4
InternVL2_5-78B ↓3 17.2
- 5 Claude 3.5 Sonnet ↑2 13.7
- 6 Qwen2-VL-7B-Instruct ↑9 12.8
- 7
InternVL2.5-38B ↓5 9.98
- 8 GPT-4o Mar 2025 ↓3 9.60
- 9
InternVL2.5-8B 7.82
- 10
InternVL2.5-4B 5.60
T5
36 models we track have a score here.
- 1 Qwen2-VL-72B-Instruct ↑5 73.0
- 2 GPT-4o Mar 2025 ↑3 73.0
- 3 Claude 3.5 Sonnet ↑4 72.3
- 4
InternVL2_5-78B ↓3 71.9
- 5
Gemini 1.5 Pro Sep 2024 ↑3 69.9
- 6
InternVL2.5-8B ↑3 69.5
- 7 Qwen2-VL-7B-Instruct ↑8 68.8
- 8
Gemini 2.0 Flash Feb 2025 ↓5 68.8
- 9
InternVL2.5-26B ↓5 68.0
- 10 MiniCPM-V-2_6 ↑6 66.8
T6
36 models we track have a score here.
- 1 Qwen2-VL-7B-Instruct ↑14 57.7
- 2
MiniCPM-V ↑25 57.4
- 3 Qwen2-VL-72B-Instruct ↑3 56.9
- 4 GLM-4V-9B ↑9 56.5
- 5
InternVL2.5-38B ↓3 56.4
- 6 MiniCPM-V-2_6 ↑10 55.7
- 7
Qwen2-VL-2B-Instruct ↑12 55.6
- 8
InternVL2_5-78B ↓7 55.5
- 9
MiniCPM-V 2.0 ↑15 54.7
- 10 GPT-4o Mar 2025 ↓5 54.6
T7
36 models we track have a score here.
- 1
QVQ-72B-Preview ↑11 77.9
- 2
Gemini 2.0 Flash Feb 2025 ↑1 77.9
- 3 Qwen2-VL-72B-Instruct ↑3 75.2
- 4
InternVL2_5-78B ↓3 70.3
- 5 Claude 3.5 Sonnet ↑2 69.0
- 6 GPT-4o Mar 2025 ↓1 69.0
- 7 Qwen2-VL-7B-Instruct ↑8 68.3
- 8
Gemini 1.5 Pro Sep 2024 67.6
- 9
InternVL2.5-38B ↓7 66.2
- 10 GLM-4V-9B ↑3 62.8
T8
36 models we track have a score here.
- 1
InternVL2_5-78B 68.7
- 2
InternVL2-8B ↑12 63.1
- 3
InternVL2.5-1B ↑14 61.5
- 4
InternVL2.5-4B ↑6 60.9
- 5
InternVL2.5-8B ↑4 60.4
- 6
InternVL2.5-38B ↓4 57.4
- 7
InternVL2.5-26B ↓3 57.0
- 8
InternVL2-26B ↑3 53.0
- 9
InternVL2-4B ↑11 48.0
- 10
QVQ-72B-Preview ↑2 45.7
T9
36 models we track have a score here.
- 1
InternVL2_5-78B 68.7
- 2
InternVL2-8B ↑12 63.1
- 3
InternVL2.5-1B ↑14 61.5
- 4
InternVL2.5-4B ↑6 60.9
- 5
InternVL2.5-8B ↑4 60.4
- 6
InternVL2.5-38B ↓4 57.4
- 7
InternVL2.5-26B ↓3 57.0
- 8
InternVL2-26B ↑3 53.0
- 9
InternVL2-4B ↑11 48.0
- 10
QVQ-72B-Preview ↑2 44.1
T10
36 models we track have a score here.
- 1
InternVL2.5-38B ↑1 44.5
- 2
InternVL2_5-78B ↓1 44.5
- 3 Qwen2-VL-72B-Instruct ↑3 43.7
- 4
Gemini 1.5 Pro Sep 2024 ↑4 43.7
- 5
InternVL2.5-26B ↓1 43.7
- 6 Claude 3.5 Sonnet ↑1 42.0
- 7
InternVL2-26B ↑4 41.2
- 8
InternVL2.5-1B ↑9 40.3
- 9
InternVL2.5-4B ↑1 40.3
- 10
InternVL2-8B ↑4 39.5
T11
36 models we track have a score here.
- 1
InternVL2.5-26B ↑3 44.8
- 2 Qwen2-VL-72B-Instruct ↑4 42.9
- 3
InternVL2.5-1B ↑14 41.0
- 4
MiniCPM-Llama3-V 2.5 ↑21 40.0
- 5
InternVL2.5-2B ↑13 40.0
- 6
DeepSeek-VL2-Tiny ↑15 40.0
- 7
QVQ-72B-Preview ↑5 40.0
- 8 GLM-4V-9B ↑5 38.1
- 9
Molmo-7B-D ↑26 38.1
- 10
InternVL2_5-78B ↓9 37.1
T12
36 models we track have a score here.
- 1
InternVL2_5-78B 76.7
- 2
InternVL2.5-4B ↑8 69.3
- 3
Gemini 2.0 Flash Feb 2025 68.4
- 4
InternVL2.5-26B 67.4
- 5 GPT-4o Mar 2025 67.4
- 6
InternVL2.5-8B ↑3 66.5
- 7
QVQ-72B-Preview ↑5 66.0
- 8
InternVL2.5-38B ↓6 64.2
- 9 Qwen2-VL-72B-Instruct ↓3 63.3
- 10 GLM-4V-9B ↑3 58.1
T13
36 models we track have a score here.
- 1 GPT-4o Mar 2025 ↑4 70.4
- 2 GLM-4V-9B ↑11 66.4
- 3 Claude 3.5 Sonnet ↑4 63.2
- 4
Gemini 2.0 Flash Feb 2025 ↓1 61.6
- 5
Gemini 1.5 Pro Sep 2024 ↑3 52.8
- 6 MiniCPM-V-2_6 ↑10 52.0
- 7 Qwen2-VL-72B-Instruct ↓1 52.0
- 8
QVQ-72B-Preview ↑4 52.0
- 9
InternVL2.5-8B 46.4
- 10
InternVL2.5-38B ↓8 44.0
T14
36 models we track have a score here.
- 1
Gemini 1.5 Pro Sep 2024 ↑7 39.0
- 2
Gemini 2.0 Flash Feb 2025 ↑1 37.3
- 3 Qwen2-VL-72B-Instruct ↑3 36.4
- 4
QVQ-72B-Preview ↑8 36.4
- 5
Molmo-7B-D ↑30 35.6
- 6
MiniCPM-Llama3-V 2.5 ↑19 33.9
- 7
InternVL2-8B ↑7 33.9
- 8
InternVL2.5-8B ↑1 33.0
- 9
InternVL2-26B ↑2 32.2
- 10 MiniCPM-V-2_6 ↑6 32.2
T15
36 models we track have a score here.
- 1 GPT-4o Mar 2025 ↑4 48.0
- 2 Qwen-VL-Chat ↑20 47.3
- 3
Qwen2-VL-2B-Instruct ↑16 47.3
- 4
InternVL2.5-2B ↑14 46.9
- 5
InternVL2.5-8B ↑4 46.9
- 6 Qwen2-VL-7B-Instruct ↑9 46.7
- 7
InternVL2.5-4B ↑3 46.5
- 8 Claude 3.5 Sonnet ↓1 46.5
- 9
InternVL2.5-38B ↓7 46.3
- 10 Qwen2-VL-72B-Instruct ↓4 46.3
T16
36 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑2 79.9
- 2
Gemini 1.5 Pro Sep 2024 ↑6 76.5
- 3
InternVL2.5-38B ↓1 73.2
- 4
InternVL2.5-26B 72.5
- 5
InternVL2.5-8B ↑4 72.5
- 6
InternVL2.5-4B ↑4 71.8
- 7 MiniCPM-V-2_6 ↑9 70.5
- 8
InternVL2-26B ↑3 65.8
- 9
InternVL2-8B ↑5 65.1
- 10
QVQ-72B-Preview ↑2 64.4
T17
36 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑6 46.7
- 2 GPT-4o Mar 2025 ↑3 46.6
- 3
DeepSeek-VL2-Tiny ↑18 46.3
- 4
InternVL2.5-4B ↑6 45.9
- 5 GLM-4V-9B ↑8 45.7
- 6
InternVL2.5-2B ↑12 45.3
- 7
Qwen2-VL-2B-Instruct ↑12 45.2
- 8
InternVL2.5-1B ↑9 45.0
- 9
InternVL2.5-26B ↓5 44.9
- 10
InternVL2_5-78B ↓9 44.9
T18
36 models we track have a score here.
- 1
InternVL2_5-78B 78.8
- 2
Gemini 2.0 Flash Feb 2025 ↑1 74.0
- 3
InternVL2.5-38B ↓1 70.5
- 4 GPT-4o Mar 2025 ↑1 69.9
- 5
InternVL2.5-26B ↓1 67.8
- 6 Qwen2-VL-72B-Instruct 66.4
- 7 GLM-4V-9B ↑6 63.0
- 8 Claude 3.5 Sonnet ↓1 60.3
- 9
InternVL2-26B ↑2 60.3
- 10
InternVL2-8B ↑4 60.3
T19
36 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑2 79.1
- 2 MiniCPM-V-2_6 ↑14 75.5
- 3
Gemini 1.5 Pro Sep 2024 ↑5 74.5
- 4 GPT-4o Mar 2025 ↑1 74.5
- 5 Claude 3.5 Sonnet ↑2 70.9
- 6 Qwen2-VL-72B-Instruct 65.5
- 7
InternVL2_5-78B ↓6 65.5
- 8 GLM-4V-9B ↑5 63.6
- 9
InternVL2.5-4B ↑1 61.8
- 10
InternVL2.5-1B ↑7 60.9
T20
36 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑2 87.8
- 2
QVQ-72B-Preview ↑10 86.4
- 3
InternVL2.5-38B ↓1 85.0
- 4 Qwen2-VL-72B-Instruct ↑2 84.3
- 5
Gemini 1.5 Pro Sep 2024 ↑3 84.3
- 6
InternVL2.5-26B ↓2 84.3
- 7
InternVL2_5-78B ↓6 83.7
- 8 GLM-4V-9B ↑5 83.0
- 9
InternVL2-26B ↑2 83.0
- 10 GPT-4o Mar 2025 ↓5 82.3
T21
36 models we track have a score here.
- 1
InternVL2_5-78B 63.6
- 2 GPT-4o Mar 2025 ↑3 63.6
- 3
InternVL2.5-38B ↓1 60.0
- 4 Claude 3.5 Sonnet ↑3 58.2
- 5 Qwen2-VL-72B-Instruct ↑1 54.5
- 6
Gemini 1.5 Pro Sep 2024 ↑2 54.5
- 7
InternVL2.5-26B ↓3 54.5
- 8
InternVL2.5-4B ↑2 54.5
- 9 Qwen2-VL-7B-Instruct ↑6 52.7
- 10 GLM-4V-9B ↑3 50.9
T22
36 models we track have a score here.
- 1 GPT-4o Mar 2025 ↑4 44.6
- 2 Qwen2-VL-7B-Instruct ↑13 44.0
- 3 Qwen-VL-Chat ↑19 43.8
- 4 Qwen2-VL-72B-Instruct ↑2 43.5
- 5 Claude 3.5 Sonnet ↑2 43.5
- 6
InternVL2.5-4B ↑4 43.5
- 7
Qwen2-VL-2B-Instruct ↑12 43.3
- 8 GLM-4V-9B ↑5 43.1
- 9
Ovis1.5-Gemma2-9B ↑19 43.1
- 10
Gemini 1.5 Pro Sep 2024 ↓2 42.8
T23
36 models we track have a score here.
- 1
InternVL2_5-78B 70.0
- 2
InternVL2.5-4B ↑8 68.3
- 3
InternVL2.5-8B ↑6 66.7
- 4
Gemini 2.0 Flash Feb 2025 ↓1 66.7
- 5 Qwen2-VL-72B-Instruct ↑1 65.8
- 6
Gemini 1.5 Pro Sep 2024 ↑2 65.8
- 7 GPT-4o Mar 2025 ↓2 65.8
- 8
QVQ-72B-Preview ↑4 65.0
- 9
InternVL2.5-38B ↓7 64.2
- 10 MiniCPM-V-2_6 ↑6 63.3
T24
36 models we track have a score here.
- 1
InternVL2_5-78B 70.8
- 2
InternVL2.5-26B ↑2 67.7
- 3
InternVL2.5-38B ↓1 67.7
- 4
Gemini 2.0 Flash Feb 2025 ↓1 67.7
- 5
InternVL2.5-8B ↑4 67.3
- 6 Claude 3.5 Sonnet ↑1 64.6
- 7
QVQ-72B-Preview ↑5 64.6
- 8 Qwen2-VL-72B-Instruct ↓2 62.5
- 9 GPT-4o Mar 2025 ↓4 61.5
- 10
Gemini 1.5 Pro Sep 2024 ↓2 58.3
T25
36 models we track have a score here.
- 1
InternVL2_5-78B 51.1
- 2 GPT-4o Mar 2025 ↑3 51.0
- 3
InternVL2.5-38B ↓1 50.9
- 4
InternVL2.5-4B ↑6 50.9
- 5
InternVL2.5-26B ↓1 50.8
- 6 Qwen2-VL-72B-Instruct 50.7
- 7 Qwen2-VL-7B-Instruct ↑8 50.6
- 8
InternVL2.5-2B ↑10 50.6
- 9
InternVL2.5-8B 50.6
- 10 Claude 3.5 Sonnet ↓3 50.3
T26
36 models we track have a score here.
- 1
InternVL2_5-78B 72.4
- 2
InternVL2.5-38B 70.5
- 3
InternVL2.5-26B ↑1 68.3
- 4
InternVL2.5-8B ↑5 68.1
- 5
InternVL2.5-4B ↑5 66.2
- 6
Gemini 2.0 Flash Feb 2025 ↓3 66.1
- 7
InternVL2-8B ↑7 65.3
- 8
InternVL2-26B ↑3 62.0
- 9 Qwen2-VL-72B-Instruct ↓3 59.7
- 10
InternVL2-4B ↑10 57.0
T27
36 models we track have a score here.
- 1 GPT-4o Mar 2025 ↑4 50.5
- 2
InternVL2.5-4B ↑8 50.4
- 3 GLM-4V-9B ↑10 50.1
- 4 Qwen2-VL-7B-Instruct ↑11 50.0
- 5
InternVL2.5-38B ↓3 50.0
- 6
Qwen2-VL-2B-Instruct ↑13 49.2
- 7
Ovis1.5-Gemma2-9B ↑21 49.0
- 8
InternVL2.5-8B ↑1 48.8
- 9
InternVL2.5-1B ↑8 48.7
- 10 Qwen2-VL-72B-Instruct ↓4 48.7
T28
36 models we track have a score here.
- 1
InternVL2_5-78B 8.16
- 2
InternVL2.5-26B ↑2 7.48
- 3
InternVL2.5-38B ↓1 6.80
- 4
MiniCPM-Llama3-V 2.5 ↑21 5.44
- 5 Claude 3.5 Sonnet ↑2 5.44
- 6
InternVL2.5-4B ↑4 5.44
- 7
Gemini 2.0 Flash Feb 2025 ↓4 5.44
- 8
InternVL2-26B ↑3 4.76
- 9 MiniCPM-V-2_6 ↑7 4.76
- 10 Qwen2-VL-7B-Instruct ↑5 4.76
T29
36 models we track have a score here.
- 1
MiniCPM-V ↑26 49.0
- 2 GLM-4V-9B ↑11 48.5
- 3 GPT-4o Mar 2025 ↑2 47.6
- 4 Qwen-VL-Chat ↑18 47.5
- 5
MiniCPM-V 2.0 ↑19 47.0
- 6
InternVL2.5-8B ↑3 47.0
- 7
InternVL2.5-2B ↑11 47.0
- 8
InternVL2.5-38B ↓6 46.8
- 9
InternVL2_5-78B ↓8 46.8
- 10
Ovis1.5-Gemma2-9B ↑18 46.7
T30
36 models we track have a score here.
- 1
QVQ-72B-Preview ↑11 15.6
- 2
InternVL2_5-78B ↓1 15.2
- 3
InternVL2.5-38B ↓1 11.5
- 4
InternVL2.5-26B 10.9
- 5
InternVL2.5-4B ↑5 9.71
- 6
InternVL2-26B ↑5 9.41
- 7
InternVL2.5-1B ↑10 8.30
- 8 Claude 3.5 Sonnet ↓1 8.22
- 9
InternVL2.5-8B 7.83
- 10
InternVL2-8B ↑4 7.51
T31
36 models we track have a score here.
- 1
InternVL2_5-78B 42.7
- 2
InternVL2-26B ↑9 41.3
- 3
InternVL2.5-38B ↓1 41.3
- 4 GLM-4V-9B ↑9 40.0
- 5
Gemini 1.5 Pro Sep 2024 ↑3 38.0
- 6
Gemini 2.0 Flash Feb 2025 ↓3 37.3
- 7 GPT-4o Mar 2025 ↓2 37.3
- 8
InternVL2.5-26B ↓4 36.7
- 9 Claude 3.5 Sonnet ↓2 34.7
- 10
InternVL2.5-4B 34.7
T32
36 models we track have a score here.
- 1 GPT-4o Mar 2025 ↑4 100.0
- 2
InternVL2_5-78B ↓1 99.3
- 3
InternVL2.5-26B ↑1 98.7
- 4 Claude 3.5 Sonnet ↑3 98.0
- 5
InternVL2.5-38B ↓3 98.0
- 6
Gemini 2.0 Flash Feb 2025 ↓3 97.3
- 7 GLM-4V-9B ↑6 96.7
- 8 Qwen2-VL-72B-Instruct ↓2 96.7
- 9
Gemini 1.5 Pro Sep 2024 ↓1 96.7
- 10
QVQ-72B-Preview ↑2 96.7
T33
36 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑2 69.8
- 2 Claude 3.5 Sonnet ↑5 62.1
- 3
MiniCPM-V ↑24 61.2
- 4
InternVL2_5-78B ↓3 61.2
- 5 GPT-4o Mar 2025 58.6
- 6
InternVL2.5-26B ↓2 55.2
- 7 GLM-4V-9B ↑6 54.3
- 8 MiniCPM-V-2_6 ↑8 53.5
- 9
InternVL2-26B ↑2 52.6
- 10
QVQ-72B-Preview ↑2 52.6
T34
36 models we track have a score here.
- 1 Qwen2-VL-72B-Instruct ↑5 49.2
- 2
InternVL2_5-78B ↓1 49.2
- 3
Gemini 2.0 Flash Feb 2025 47.5
- 4 MiniCPM-V-2_6 ↑12 45.0
- 5 Claude 3.5 Sonnet ↑2 42.5
- 6
Qwen2-VL-2B-Instruct ↑13 42.5
- 7 GPT-4o Mar 2025 ↓2 42.5
- 8
InternVL2.5-4B ↑2 41.7
- 9
InternVL2.5-38B ↓7 40.8
- 10
DeepSeek-VL2-Tiny ↑11 40.8
T35
36 models we track have a score here.
- 1 Qwen-VL-Chat ↑21 47.5
- 2 GPT-4o Mar 2025 ↑3 46.3
- 3 GLM-4V-9B ↑10 45.9
- 4 llava-v1.5-7b ↑30 45.9
- 5
MiniCPM-V ↑22 45.8
- 6 Qwen2-VL-7B-Instruct ↑9 45.7
- 7 Claude 3.5 Sonnet 45.6
- 8
Ovis1.6-Gemma2-9B ↑15 45.2
- 9
InternVL2.5-4B ↑1 44.9
- 10
Qwen2-VL-2B-Instruct ↑9 44.7
T36
36 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑2 94.6
- 2 GPT-4o Mar 2025 ↑3 94.6
- 3
Gemini 1.5 Pro Sep 2024 ↑5 93.8
- 4 Claude 3.5 Sonnet ↑3 92.9
- 5 GLM-4V-9B ↑8 91.1
- 6
InternVL2_5-78B ↓5 91.1
- 7 Qwen2-VL-72B-Instruct ↓1 90.2
- 8
InternVL2.5-26B ↓4 90.2
- 9 Qwen-VL-Chat ↑13 88.4
- 10
InternVL2-26B ↑1 88.4
T37
36 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑2 70.5
- 2
Gemini 1.5 Pro Sep 2024 ↑6 57.1
- 3
InternVL2-4B ↑17 50.8
- 4
InternVL2_5-78B ↓3 48.2
- 5
InternVL2.5-38B ↓3 39.5
- 6
InternVL2.5-8B ↑3 38.2
- 7 Claude 3.5 Sonnet 36.4
- 8 Qwen2-VL-72B-Instruct ↓2 30.6
- 9
InternVL2.5-4B ↑1 29.0
- 10
InternVL2-26B ↑1 25.1
T38
36 models we track have a score here.
- 1
Molmo-7B-O ↑35 100.0
- 2
Molmo-7B-D ↑33 100.0
- 3
MiniCPM-V ↑24 99.5
- 4 llava-v1.5-7b ↑30 99.5
- 5
LLaVA-v1.5-13B ↑26 99.3
- 6 llava-v1.6-vicuna-7b ↑24 98.8
- 7 llava-v1.6-mistral-7b ↑26 98.7
- 8
MiniCPM-V 2.0 ↑16 98.7
- 9
Ovis1.5-Gemma2-9B ↑19 98.6
- 10 llava-v1.6-vicuna-13b ↑22 98.5
T39
36 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑2 81.5
- 2
InternVL2_5-78B ↓1 80.6
- 3
Gemini 1.5 Pro Sep 2024 ↑5 80.6
- 4 Qwen2-VL-72B-Instruct ↑2 77.2
- 5
InternVL2.5-38B ↓3 76.0
- 6
InternVL2.5-26B ↓2 73.7
- 7
InternVL2-8B ↑7 70.3
- 8
InternVL2.5-8B ↑1 70.2
- 9
InternVL2-26B ↑2 69.9
- 10
InternVL2.5-4B 67.1
T40
36 models we track have a score here.
- 1
InternVL2_5-78B 90.7
- 2
InternVL2.5-26B ↑2 90.1
- 3
InternVL2.5-38B ↓1 89.3
- 4
InternVL2.5-8B ↑5 89.2
- 5 Qwen2-VL-7B-Instruct ↑10 87.8
- 6
InternVL2.5-1B ↑11 87.5
- 7
InternVL2.5-4B ↑3 87.4
- 8
InternVL2-8B ↑6 87.2
- 9
InternVL2-26B ↑2 86.2
- 10
InternVL2.5-2B ↑8 86.2
T41
36 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑2 80.7
- 2
Gemini 1.5 Pro Sep 2024 ↑6 78.9
- 3
InternVL2_5-78B ↓2 77.2
- 4 Qwen2-VL-72B-Instruct ↑2 72.8
- 5
InternVL2.5-38B ↓3 72.3
- 6
InternVL2.5-26B ↓2 68.8
- 7
InternVL2-8B ↑7 65.6
- 8
InternVL2.5-8B ↑1 65.3
- 9
InternVL2-26B ↑2 64.9
- 10
InternVL2-4B ↑10 62.0
T42
36 models we track have a score here.
- 1
InternVL2_5-78B 66.9
- 2
Gemini 2.0 Flash Feb 2025 ↑1 62.0
- 3
InternVL2.5-38B ↓1 61.5
- 4
Gemini 1.5 Pro Sep 2024 ↑4 59.8
- 5
InternVL2.5-26B ↓1 56.3
- 6
InternVL2.5-8B ↑3 53.8
- 7 Qwen2-VL-72B-Instruct ↓1 53.4
- 8
InternVL2-8B ↑6 52.1
- 9
InternVL2-26B ↑2 51.2
- 10
InternVL2.5-4B 50.0
T43
36 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑2 60.5
- 2
Gemini 1.5 Pro Sep 2024 ↑6 58.6
- 3 Claude 3.5 Sonnet ↑4 41.7
- 4
InternVL2_5-78B ↓3 38.9
- 5 Qwen2-VL-72B-Instruct ↑1 37.1
- 6
InternVL2.5-26B ↓2 36.4
- 7 GPT-4o Mar 2025 ↓2 33.9
- 8
InternVL2-26B ↑3 33.9
- 9
InternVL2.5-8B 32.3
- 10
InternVL2.5-4B 32.2
T44
36 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑2 28.5
- 2
Gemini 1.5 Pro Sep 2024 ↑6 27.9
- 3
InternVL2_5-78B ↓2 14.4
- 4 Qwen2-VL-72B-Instruct ↑2 14.1
- 5
InternVL2.5-26B ↓1 13.5
- 6 Claude 3.5 Sonnet ↑1 12.9
- 7
InternVL2-26B ↑4 12.3
- 8
InternVL2.5-4B ↑2 12.1
- 9
InternVL2.5-8B 11.5
- 10
InternVL2-4B ↑10 11.0
T45
36 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑2 44.3
- 2
Gemini 1.5 Pro Sep 2024 ↑6 42.0
- 3
InternVL2_5-78B ↓2 30.2
- 4
InternVL2.5-26B 27.9
- 5 Qwen2-VL-72B-Instruct ↑1 27.1
- 6
InternVL2-26B ↑5 26.5
- 7 Claude 3.5 Sonnet 25.8
- 8
InternVL2.5-4B ↑2 25.6
- 9
InternVL2-4B ↑11 25.1
- 10
InternVL2.5-8B ↓1 24.8
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: The MCS-Bench paper
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official MCS-Bench 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
8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 20 months ago, in March 2025, for GPT-4o. Every benchmark
Catalog record 2 pass
- Pass Task access The tasks are private.
- Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
- Pass Independent runs Scores from the benchmark’s paper (37 entries).
- Pass Models scored 36 models we track, plus 1 other entry on its leaderboard.
- Warning Newest score The newest came 20 months ago, in March 2025, for GPT-4o.
- Note Frontier models None of the top 10 has a score. The strongest tested is Gemini 2.0 Flash (#147).
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on github.com: 36 rows for models we track and 1 other entry 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.