Skip to main content
leaderboard
.cn
Toggle Sidebar
Search models or organizations…
⌘K
Explore
Frontier Models
Historical Trends
Local Models
Compare Models
Ranking Studio
Library
Models
Pricing
Organizations
Benchmarks
Data sources
Settings
Toggle Sidebar
leaderboard
.cn
Loading page content
Feedback
Loading page content
Back to models
DeepSeek-V3 Base
Open
Share
Add to comparison
Overview
Evaluations
Evaluation overview
Overall rank
#127
60.2 points · 180 models
Open-weight rank
#62
65.6 points · 109 models
Data coverage
15 items
1 source · updated through 2025-05-14
Domain performance
Coding
#125 / 160
62.1
Research
#124 / 174
70.1
Evaluation highlights
All evaluations
CRUX-O
Coding · Qwen3 Technical Report
76.6%
3
/ 25
MMMLU
Reasoning · Qwen3 Technical Report
85.88%
8
/ 40
MGSM
Multilingual · Qwen3 Technical Report
82.68%
3
/ 25
MATH
Reasoning · Qwen3 Technical Report
62.62%
8
/ 30
MBPP
Coding · Qwen3 Technical Report
74.2%
7
/ 25
BBH
Reasoning · Qwen3 Technical Report
86.22%
9
/ 30
Basic information
Context (input / output)
131K / 8K
Parameters
671B / 37B active
Release date
2024-12
License
MIT
Input modalities
Text
Output modalities
Text
Active parameters
37B
Related links
Official site
HuggingFace
Version relationships
DeepSeek-V3 family
DeepSeek-V3
2024-12 · Same series
DeepSeek-V3 Base
Current
2024-12 · Current model
DeepSeek-V3-0324
2025-03 · Same series
DeepSeek-V3.2
2025-12 · Same series