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-V4-Flash
Open
Share
Add to comparison
Tool use · Reasoning
Overview
Deployment
Pricing
Evaluations
Evaluation overview
Overall rank
#41
77.1 points · 180 models
Open-weight rank
#8
86.6 points · 109 models
Data coverage
73 items
5 sources · updated through 2026-07-16
Domain performance
Coding
#39 / 160
78
Research
#39 / 174
87.6
Agents
#38 / 125
76.2
Long context
#53 / 82
70
Evaluation highlights
All evaluations
τ³-Bench
Agents · Nemotron-3-Ultra Technical Report
73.7%
1
/ 15
LiveCodeBench-v6
Reasoning · Nemotron-3-Ultra Technical Report
90.9%
4
/ 49
SWE-bench Verified
Coding · Hy3 Hugging Face README
79%
15
/ 60
LongBench v2
Long context · Nemotron-3-Ultra Technical Report
57%
11
/ 22
IMOAnswerBench
Reasoning · Nemotron-3-Ultra Technical Report
91.1%
4
/ 35
IFBench (prompt loose)
Reasoning · Nemotron-3-Ultra Technical Report
82%
1
/ 7
Basic information
Context (input / output)
1M / 393K
Parameters
284B / 13B active
Release date
2026-04
License
MIT
Input modalities
Text
Output modalities
Text
Active parameters
13B
Availability
DeepSeek
Full pricing
Input
≈ $0.1429 /1M
Output
≈ $0.2857 /1M
vLLM · FP8
Deployment options
FP8 · 8 × 141 GB VRAM
Related links
Official site
HuggingFace
ModelScope
Version relationships
DeepSeek-V4 family
DeepSeek-V4-Flash
Current
2026-04 · Current model
DeepSeek-V4-Flash Base
2026-04 · Same series
DeepSeek-V4-Pro
2026-04 · Same series
DeepSeek-V4-Pro Base
2026-04 · Same series
Runtime modes
Thinking