DeepSeek V3
DeepSeek DeepSeek V3 — pricing, 164k context window, API cost calculator and alternatives.
By DeepSeek
Specs
- Provider
- DeepSeek
- Slug
- deepseek/deepseek-chat
- Capabilities
- tools, json_mode
Pricing freshness
- Tier
- standard
- Currency
- USD
- As of
- 2026-05-08 17:08 UTC
Pricing history
Tracking DeepSeek V3 pricing since 2026-05-08. We'll plot the chart here once it changes.
Quickstart — call DeepSeek V3 from your app
curl https://api.deepseek.com/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $DEEPSEEK_API_KEY" \
-d '{
"model": "deepseek-chat",
"messages": [{"role": "user", "content": "Hello!"}]
}'Official docs: https://api-docs.deepseek.com/
# pip install openai
from openai import OpenAI
client = OpenAI(
base_url="https://api.deepseek.com/v1",
api_key=os.environ["DEEPSEEK_API_KEY"],
)
resp = client.chat.completions.create(
model="deepseek-chat",
messages=[{"role": "user", "content": "Hello!"}],
)
print(resp.choices[0].message.content)Official docs: https://api-docs.deepseek.com/
// npm install openai
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api.deepseek.com/v1",
apiKey: process.env.DEEPSEEK_API_KEY,
});
const resp = await client.chat.completions.create({
model: "deepseek-chat",
messages: [{ role: "user", content: "Hello!" }],
});
console.log(resp.choices[0].message.content);Official docs: https://api-docs.deepseek.com/
Related models
Similar capabilities, context window, and price tier — drawn from across the catalog so you can compare alternatives in one click.
Frequently asked questions
What is DeepSeek V3?
DeepSeek V3 is a large language model API from DeepSeek with a 164k-token context window. It costs $0.32 per 1M input tokens and $0.89 per 1M output tokens.
How much does DeepSeek V3 cost?
DeepSeek V3 is priced at $0.32 per 1M input tokens and $0.89 per 1M output tokens via the DeepSeek API. A 50/50 input/output workload of 1M total tokens costs about $0.605.
What is the context window of DeepSeek V3?
DeepSeek V3 supports up to 164k tokens of context per request — roughly 328 pages of English text or 20480 lines of code at a typical density.
Does DeepSeek V3 support vision, tool use, or JSON mode?
DeepSeek V3 supports tool/function calling and structured JSON mode. It does not support image input (vision).
Who makes DeepSeek V3?
DeepSeek V3 is built and operated by DeepSeek. Pricing, context window, and capabilities on this page are refreshed nightly from DeepSeek's public catalog.
Can I self-host DeepSeek V3?
DeepSeek V3 is API-only — its weights are not publicly distributed by DeepSeek, so it cannot be self-hosted today.