DeepSeek Models

DeepSeek's V4 family pairs very large context windows (up to 1M tokens) with aggressive cost efficiency, making it a popular default for coding assistants and high-volume pipelines. AI Gateway exposes both current tiers below.

Available Models

Live per-model credit pricing is on the Models page.

Quick Start

Python (OpenAI SDK)

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://your-domain.com/api/ai/v1"
)

response = client.chat.completions.create(
    model="deepseek-v4-pro",
    messages=[
        {"role": "system", "content": "You are a senior backend engineer."},
        {"role": "user", "content": "Implement a sliding-window rate limiter."}
    ]
)
print(response.choices[0].message.content)

Streaming

stream = client.chat.completions.create(
    model="deepseek-v4-flash",
    messages=[{"role": "user", "content": "Explain vector databases."}],
    stream=True,
)
for chunk in stream:
    print(chunk.choices[0].delta.content or "", end="")

Replace your-domain.com with your AI Gateway site domain. Get your API key from the dashboard after sign-up.

Why through AI Gateway

• One key for DeepSeek + all other providers — swap models per request
• 1M context for repository-scale coding tasks
• Credit billing — no provider account, no minimum spend

Related Pages

• Live model pricing
• API reference

Why through AI Gateway

• One unified interface: call every provider through a single OpenAI-compatible base URL
• One key: no per-vendor accounts; credits are shared and never expire
• Per-token billing: cache hits billed at steep discounts
• Flexible switching: change one string to move between models or vendors

About this data

Specs and prices come from the platform live catalog; performance figures from vendor or third-party publications. Updated 2026-10-13.

Why through AI Gateway

• One unified interface: call every provider through a single OpenAI-compatible base URL
• One key: no per-vendor accounts; credits are shared and never expire
• Per-token billing: cache hits billed at steep discounts
• Flexible switching: change one string to move between models or vendors

Capabilities

7 shipping models, context windows from 66K to 1M tokens; tiers: balanced ×3, flagship ×2, light ×2.

Pricing & billing

Per-token credit billing; cache hits at steep discounts. Live rates on the Models pricing page — numbers intentionally not fixed here.

FAQ

How large is the context window of DeepSeek?

1,000,000 Token (1M)

How is it billed?

Per-token credit billing with cache-hit discounts. Live rates on the Models page.

Do I need a provider account?

No. One AI Gateway API key covers this model and every other model on the site.

About this data

Specs and prices come from the platform live catalog; performance figures from vendor or third-party publications. Updated 2026-10-13.