Home/Models/DeepSeek V3

DeepSeek V3

Open Source
by DeepSeek· Released 2024-12
Official Docs ↗

DeepSeek V3 shocked the AI industry by achieving GPT-4o-level performance at a fraction of the cost. At $0.27/1M input tokens with open weights, it offers the best value proposition of any frontier-class model in 2025.

Input Price
$0.27/1M
per 1M tokens
Output Price
$1.10/1M
per 1M tokens
Context Window
131K tokens
max tokens
ELO Score
1320
LMSYS Arena

Pricing Breakdown

VolumeInput CostOutput CostCombined (50/50)
1,000 tokens$0.0003$0.0011$0.0007
10,000 tokens$0.0027$0.0110$0.0069
100,000 tokens$0.0270$0.1100$0.0685
1,000,000 tokens$0.2700$1.1000$0.6850

Strengths

  • Frontier-level quality at 10x lower cost than GPT-4o
  • Open source weights available
  • Excellent coding performance
  • Strong on English and Chinese tasks

Weaknesses

  • China-based company may raise data privacy concerns for some
  • Smaller ecosystem than US providers

Best For

codinglow costreasoningdata extraction

Compare DeepSeek V3 With

DeepSeek V3 vs o3
OpenAI
DeepSeek V3 vs DeepSeek R1
DeepSeek
DeepSeek V3 vs o1
OpenAI
DeepSeek V3 vs Gemini 2.0 Flash
Google
DeepSeek V3 vs Claude Sonnet 4.6
Anthropic
DeepSeek V3 vs Claude 3.5 Sonnet
Anthropic

Frequently Asked Questions

How much does DeepSeek V3 cost?

DeepSeek V3 costs $0.27/1M per 1M input tokens and $1.10/1M per 1M output tokens.

What is DeepSeek V3's context window?

DeepSeek V3 has a context window of 131K tokens, which means it can process up to 131,072 tokens in a single request.

Is DeepSeek V3 open source?

Yes, DeepSeek V3 is open source. The model weights are publicly available and can be self-hosted.

What is DeepSeek V3 best used for?

DeepSeek V3 is best suited for: coding, low-cost, reasoning, data-extraction. Frontier-level quality at 10x lower cost than GPT-4o.

What is DeepSeek V3's ELO score?

DeepSeek V3 has an ELO score of 1320 on the LMSYS Chatbot Arena leaderboard, placing it among top-tier models.