DeepSeek VL2 Small
Multimodal
Zero-eval
#2MMBench-V1.1
#2MME
#3MMT-Bench
by DeepSeek
About
DeepSeek VL2 Small is a multimodal language model developed by DeepSeek. It achieves strong performance with an average score of 69.6% across 14 benchmarks. It excels particularly in DocVQA (92.3%), ChartQA (84.5%), TextVQA (83.4%). As a multimodal model, it can process and understand text, images, and other input formats seamlessly. Released in 2024, it represents DeepSeek's latest advancement in AI technology.
Timeline
AnnouncedDec 13, 2024
ReleasedDec 13, 2024
Specifications
Capabilities
Multimodal
License & Family
License
deepseek
Benchmark Performance Overview
Performance metrics and category breakdown
Overall Performance
14 benchmarks
Average Score
69.6%
Best Score
92.3%
High Performers (80%+)
6Top Categories
vision
74.9%
general
68.9%
roleplay
62.9%
math
60.7%
Benchmark Performance
Top benchmark scores with normalized values (0-100%)
Ranking Across Benchmarks
Position relative to other models on each benchmark
DocVQA
Rank #17 of 26
#14Nova Lite
92.4%
#15GPT-4o
92.8%
#16Phi-4-multimodal-instruct
93.2%
#17DeepSeek VL2 Small
92.3%
#18Pixtral-12B
90.7%
#19Llama 3.2 90B Instruct
90.1%
#20DeepSeek VL2 Tiny
88.9%
ChartQA
Rank #15 of 24
#12Qwen2.5-Omni-7B
85.3%
#13Llama 3.2 90B Instruct
85.5%
#14GPT-4o
85.7%
#15DeepSeek VL2 Small
84.5%
#16Llama 3.2 11B Instruct
83.4%
#17Pixtral-12B
81.8%
#18Phi-3.5-vision-instruct
81.8%
TextVQA
Rank #5 of 15
#2DeepSeek VL2
84.2%
#3Qwen2.5-Omni-7B
84.4%
#4Qwen2.5 VL 7B Instruct
84.9%
#5DeepSeek VL2 Small
83.4%
#6Nova Pro
81.5%
#7DeepSeek VL2 Tiny
80.7%
#8Nova Lite
80.2%
OCRBench
Rank #5 of 7
#2Phi-4-multimodal-instruct
84.4%
#3Qwen2.5 VL 7B Instruct
86.4%
#4Qwen2-VL-72B-Instruct
87.7%
#5DeepSeek VL2 Small
83.4%
#6DeepSeek VL2
81.1%
#7DeepSeek VL2 Tiny
80.9%
MMBench
Rank #5 of 7
#2Phi-3.5-vision-instruct
81.9%
#3Qwen2.5 VL 7B Instruct
84.3%
#4Phi-4-multimodal-instruct
86.7%
#5DeepSeek VL2 Small
80.3%
#6DeepSeek VL2
79.6%
#7DeepSeek VL2 Tiny
69.2%
All Benchmark Results for DeepSeek VL2 Small
Complete list of benchmark scores with detailed information
DocVQA DocVQA benchmark | vision | multimodal | 0.92 | 92.3% | Self-reported |
ChartQA ChartQA benchmark | general | multimodal | 0.84 | 84.5% | Self-reported |
TextVQA TextVQA benchmark | vision | multimodal | 0.83 | 83.4% | Self-reported |
OCRBench OCRBench benchmark | general | text | 0.83 | 83.4% | Self-reported |
MMBench MMBench benchmark | general | text | 0.80 | 80.3% | Self-reported |
AI2D AI2D benchmark | general | text | 0.80 | 80.0% | Self-reported |
MMBench-V1.1 MMBench-V1.1 benchmark | general | text | 0.79 | 79.3% | Self-reported |
InfoVQA InfoVQA benchmark | vision | multimodal | 0.76 | 75.8% | Self-reported |
RealWorldQA RealWorldQA benchmark | general | text | 0.65 | 65.4% | Self-reported |
MMT-Bench MMT-Bench benchmark | roleplay | text | 0.63 | 62.9% | Self-reported |
Showing 1 to 10 of 14 benchmarks