Up to 30% Off DeepSeek Models on LLM Gateway

DeepSeek models are already among the cheapest high-performance LLMs on the market. Now they're even cheaper on LLM Gateway.

Through our provider partnerships, we've secured exclusive discounts of up to 30% on DeepSeek V3.2, V3.1, and R1 — applied automatically to every API request. No promo codes. No minimum spend. No catch.

Discounted DeepSeek Models

DeepSeek V3.2

The latest and most capable DeepSeek model. Unified chat and reasoning in a single endpoint with a 163K context window.

Best price: $0.182/M input via Canopywave — 35% cheaper than the official DeepSeek API.

Try DeepSeek V3.2 in the Playground

DeepSeek V3.1

The previous generation, still available for workloads that depend on it.

Try DeepSeek V3.1 in the Playground

DeepSeek R1 (0528)

The May 2025 reasoning model, available through Nebius.

Try DeepSeek R1 in the Playground

How the Discounts Work

LLM Gateway negotiates volume-based pricing with infrastructure providers and passes the savings directly to you. When you send a request for a DeepSeek model, our smart routing engine:

1. Identifies all available providers for that model
2. Applies any negotiated discounts to each provider's base price
3. Routes to the cheapest option that supports your request's features (tool calling, JSON mode, etc.)
4. Falls back automatically if the primary provider is unavailable

You pay the discounted rate. We charge zero platform markup on top.

Stacking Discounts: DeepSeek Cache + LLM Gateway Partner Pricing

DeepSeek's official API already offers a 90% discount on cached input tokens — $0.028/M instead of $0.28/M for prompt cache hits. When you access DeepSeek through LLM Gateway's discounted providers, you can stack both savings:

Even with cache hits factored in, LLM Gateway's discounted providers offer competitive pricing — and you get automatic failover, usage analytics, and multi-provider routing on top.

Cost Comparison: 1M Requests

To put the discounts in perspective, here's what 1 million typical requests (1K input tokens + 200 output tokens each) would cost:

That's over $1,500 saved per year on a single model at moderate volume. Scale up to 10M requests/day and you're looking at five-figure annual savings.

Calculate your exact savings with our Token Cost Calculator

Why We Offer Discounts

Most API gateways add a markup on top of provider pricing. We don't. LLM Gateway is funded through optional paid plans (Pro and Enterprise) that offer features like extended data retention, provider key management, and priority support.

For API routing and model access, you pay exactly what the provider charges — or less, when we've negotiated a discount. It's that simple.

More Discounted Providers on LLM Gateway

DeepSeek isn't the only provider where we've secured discounts. Here's a snapshot of our current partner pricing:

• Alibaba Cloud: 20% off all 26 Qwen models — read the announcement
• Canopywave: 30% off DeepSeek, OpenAI, Anthropic, Moonshot, and MiniMax models
• Bytedance: 10% off Doubao and select hosted models
• ZAI: 10-30% off across their model catalog

Browse all discounted models

Get Started

Start using discounted DeepSeek models in under a minute:

1. Sign up free — no credit card required
2. Use our OpenAI-compatible API endpoint:

1curl https://api.llmgateway.io/v1/chat/completions \
2  -H "Authorization: Bearer YOUR_API_KEY" \
3  -H "Content-Type: application/json" \
4  -d '{
5    "model": "deepseek/deepseek-v3.2",
6    "messages": [{"role": "user", "content": "Hello!"}]
7  }'

1curl https://api.llmgateway.io/v1/chat/completions \
2  -H "Authorization: Bearer YOUR_API_KEY" \
3  -H "Content-Type: application/json" \
4  -d '{
5    "model": "deepseek/deepseek-v3.2",
6    "messages": [{"role": "user", "content": "Hello!"}]
7  }'

3. Calculate your savings with our Token Cost Calculator

Smart routing picks the cheapest discounted provider automatically. No configuration needed.

Browse discounted models | Try DeepSeek V3.2 | Token Cost Calculator | Get started free