GPU/Deploy AI Models/DeepSeek V3 0324
March 2024 Release

DeepSeek V3 0324

March 2024 release of DeepSeek V3 with 671B parameters in MoE architecture. Enhanced reasoning, coding, and multilingual capabilities with 64K context length.

Model Specifications

Parameters671B (37B active)
ArchitectureMixture of Experts (MoE)
Context Length64K tokens
Experts256 total, 8 active
LanguagesEnglish, Chinese, Code
LicenseDeepSeek License

Pricing Options

Serverless API

Pay per token with auto-scaling

₹20 input /1M tokens
₹40 output /1M tokens
  • Auto-scaling
  • No minimum
  • 99.9% uptime
  • Rate limits apply
Recommended

Dedicated Instance

Reserved GPU for consistent performance

₹400/hour
  • 8x H100 GPUs
  • No rate limits
  • Fine-tuning support
  • Private deployment

Recommended Hardware

8x

NVIDIA H100

640 GB total

Recommended for full model

4x

NVIDIA H200

564 GB total

Premium performance option

Use Cases

Code Generation

Generate high-quality code across multiple programming languages.

Mathematical Reasoning

Solve complex mathematical problems with step-by-step solutions.

Multilingual Tasks

Excellent performance in English and Chinese language tasks.

Long Context Tasks

Process and analyze documents up to 64K tokens.

Ready to Deploy DeepSeek V3 0324?

Start with our serverless API or deploy on dedicated GPUs.