The AI arena is free today

Open Superagent
API Provider3 active models1 organization

DeepSeek: API pricing, speed & models

DeepSeek hosts 3 active AI models, with input pricing from $0.28 per 1M tokens, with median throughput of 46 characters/sec, with 100% success rate over 7 days. Compare DeepSeek's API speed, pricing, and reliability against other inference providers.

First-party only
Median throughput46 c/s
Catalog latency300ms
Success rate (7d)100%
From$0.28 /M tok

Catalog

Type
Price
5 models
Model
DeepSeek-V4.1-Flashcache pricing
DeepSeek-V4.1-Flashcache pricing
At a glance

DeepSeekpricing, performance & catalog

The citable facts about DeepSeek's 3 models — sourced from provider APIs and refreshed continuously.

Lowest price
DeepSeek-V3.2 (Non-thinking) at $0.280 per 1M input tokens
Highest median throughput
DeepSeek-V4-Pro-0813 at 46 chars/s
Largest context
DeepSeek-V4.1-Flash at 1.0M tokens
Catalog
3 active models from 1 organization

Fastest median throughput

  1. 1DeepSeek-V4-Pro-081346 c/s

FAQ

Common questions about DeepSeek.

What is DeepSeek?

DeepSeek is an API provider that hosts large language models. Active models: 3; From (input): $0.28 / 1M tok; Median throughput: 46 c/s; Success rate (7d): 100%.

How many models does DeepSeek offer?

DeepSeek currently serves 3 active models out of 12 historical offerings on LLM Stats.

What is DeepSeek's API pricing?

DeepSeek input pricing starts from $0.28 per 1M tokens, with the most expensive offering at $1.32 per 1M tokens. See the Pricing tab above for the full per-model breakdown.

How fast is DeepSeek?

DeepSeek delivers a median throughput of 46 characters per second across its model catalog. See the Models tab for per-model throughput and TTFT breakdowns.

Is DeepSeek reliable?

DeepSeek has a 100% success rate across 5 API calls in the last 7 days, with a 0% error rate.

Does DeepSeek support function calling?

Yes. 5 of 3 models on DeepSeek support function calling (tool use). The Capabilities tab lists which specific models accept tool definitions.

Does DeepSeek support JSON mode and structured output?

Yes. 5 of 3 models on DeepSeek support structured output (JSON mode / schema-constrained generation). The Capabilities tab shows which specific models accept response_format or json_schema parameters.

Does DeepSeek offer batch inference?

Yes. 1 of 3 models on DeepSeek support batch inference for cheaper, asynchronous workloads.

Does DeepSeek support multimodal models?

Yes. DeepSeek's catalog includes 1 vision-capable models. See the Models and Capabilities tabs for the full per-model breakdown.

Whose models does DeepSeek host?

DeepSeek hosts models from DeepSeek. See the Models tab for the full catalog grouped by creator.

How do I start using DeepSeek?

Sign up at https://deepseek.com/ to get an API key, then call DeepSeek's API directly from your application. Most clients work out of the box by pointing the OpenAI SDK at DeepSeek's base URL with your key. Use the Models and Pricing tabs above to pick the right model for your latency, cost, and context-window requirements.