DeepSeek v3 0324
DeepSeekBalanced
ThinkingTool UseStructured Output
About this model
DeepSeek V3, a 685B-parameter, mixture-of-experts model, is the latest iteration of the flagship chat model family from the DeepSeek team. It succeeds the [DeepSeek V3](/deepseek/deepseek-chat-v3) model and performs really well on a variety of tasks.
Performance Tier
Balanced
DeepSeek v3 0324 is a balanced model from DeepSeek : strong performance at a reasonable price.
Strong cost-performance ratio. Reliable for most professional use cases without premium pricing.
Pricing
This model is included in Elosia plans| Type | per 1M tokens |
|---|---|
| Input (prompt) | $0.200 |
| Output (completion) | $0.770 |
| Cache read | $0.135 |
Capabilities
Context Length164K
Max Output Tokens—
TokenizerDeepSeek
Inputtext
Outputtext
Release DateMarch 24, 2025
Benchmarks
General Intelligence
MMLU
87.1%
MMLU-Pro
81.2%
Mathematics
MATH-500
90.2%
Programming
HumanEval
86.4%
Recommended Use Cases
CodingMathematicsAnalysis
Strengths
- Excellent value — open model competing with top proprietary models
- Strong mathematical reasoning (MATH-500 90.2%)
- MoE architecture for efficient compute usage
Limitations
- Earlier version — DeepSeek v3.2 recommended instead
- Limited multimodal capabilities
Resources
This model may use your data for training
