DeepSeek V3 0324
DeepSeekDeepSeekOpen WeightMIT + Model License (Commercial use allowed)
Description
A powerful Mixture-of-Experts (MoE) language model with 671B total parameters (37B activated per token). Features Multi-head Latent Attention (MLA), auxiliary-loss-free load balancing, and multi-token prediction training. Pre-trained on 14.8T tokens with strong performance in reasoning, math, and code tasks.
Release Date
2025-03-25
Parameters
671.0B
Context Length
164K
Modalities
text
Capability Radar
34
general
39
coding
54
reasoning
43
scienceest.
50
agents
0
multimodal
Science uses a reasoning proxy when dedicated science benchmarks are unavailable.
Rankings
| Domain | #Rank | Score | Source |
|---|---|---|---|
| Code Ranking | 233 | 41.0 | AA |
| General Ranking | 225 | 46.0 | AA |
| Math Reasoning | 165 | 54.0 | AA |
| Science | 254 | 44.0 | AA |
Benchmark Scores (LLM Stats)
Biology
GPQA
68.4%SR
Code
LiveCodeBench
49.2%SR
Finance
MMLU-Pro
81.2%SR
Math
MATH-500
94.0%SR
AIME 2024
59.4%SR
AA Evaluation Indices
Math Index41.0
Intelligence Index15.7
Math 5000.9
Mmlu Pro0.8
Gpqa0.7
Aime0.5
Tau20.5
Ifbench0.4
Aime 250.4
Lcr0.4
Livecodebench0.4
Scicode0.4
Terminalbench Hard0.2
Hle0.1
LLM Stats Category Scores
Language80
Legal80
Math80
Finance80
Healthcare80
Physics70
Reasoning70
General70
Biology70
Chemistry70
Code50
Pricing
Input Price$1.195 / 1M tokens
Output Price$1.25 / 1M tokens
Blended Price (3:1)$1.209 / 1M tokens
Cache Read Price$0.135 / 1M tokens
Speed
Tokens/sec0.0
Time to First Token0.00s
Time to Answer0.00s
Provider Price Ranking
Provider Price Ranking
14 providers
Cheapest: NovitaMost Expensive: Synthetic
ProviderInputOutput
1NovitaCheapest
$0
$0
2submodel
$0.2
$0.8
3OpenRouter
$0.2
$0.77
4Kilo Gateway
$0.2
$0.77
5NanoGPT
$0.25
$0.7
6Alibaba (China)
$0.25
$1
7Meganova
$0.25
$0.88
8NovitaAI
$0.27
$1.12
9Jiekou.AI
$0.28
$1.14
10Cortecs
$0.551
$1.654
11Azure Cognitive Services
$1.14
$4.56
12Azure
$1.14
$4.56
13DeepSeekPRIMARY
$1.195
$1.25
14Synthetic
$1.2
$1.2
Compare pricing across different API providers for this model.