DeepSeek V3

New

State-of-the-art MoE model with exceptional reasoning capabilities.

Parameters

671B MoE

Context Length

64K

About this model

DeepSeek V3 is a massive 671B parameter Mixture-of-Experts model that achieves top-tier performance while remaining cost-efficient. It uses sparse activation, meaning only a fraction of parameters are used for each token, resulting in faster inference and lower costs compared to dense models of similar capability.

Capabilities

Advanced ReasoningMath & LogicCode GenerationEfficient InferenceLong Context

Use Cases

Complex Analysis
Mathematical Problems
Code Development
Research

Model Details

Provider: DeepSeek
Model ID: deepseek-ai/DeepSeek-V3
Parameters: 671B MoE
Context Length: 64K tokens
Category: chat

API Usage

Use the DOS API to integrate DeepSeek V3into your applications. Our API is compatible with OpenAI's client libraries for easy migration.

Model ID

deepseek-ai/DeepSeek-V3

Python

python

from dos import DOS

client = DOS()

response = client.chat.completions.create(
    model="deepseek-ai/DeepSeek-V3",
    messages=[
        {"role": "user", "content": "Hello, how are you?"}
    ]
)

print(response.choices[0].message.content)

cURL

bash

curl https://api.dos.ai/v1/chat/completions \
  -H "Authorization: Bearer $DOS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-ai/DeepSeek-V3",
    "messages": [
      {"role": "user", "content": "Hello, how are you?"}
    ]
  }'

Node.js

javascript

import DOS from 'dos-ai';

const client = new DOS();

const response = await client.chat.completions.create({
  model: "deepseek-ai/DeepSeek-V3",
  messages: [
    { role: "user", content: "Hello, how are you?" }
  ]
});

console.log(response.choices[0].message.content);

View full API reference

Related Models

Qwen

Qwen 3.5 35B-A3B

Ultra-efficient MoE model — 35B total, 3B active parameters. Fast inference at near-8B cost with 70B-class quality.

$0.50 / 1M tokens

Llama 3.3 70B

High-performance multilingual LLM optimized for dialogue and instruction following.

$0.88 / 1M tokens

Llama 3.1 405B

The largest and most capable Llama model for complex reasoning and generation tasks.

$3.50 / 1M tokens