Learn

Plain-English guides to the DeepSeek API, R1 reasoning, and the broader OpenAI-compatible LLM ecosystem. No fluff, no affiliate spam.

DeepSeek basics

Cache-hit vs cache-miss input tokens, output pricing, and why DeepSeek can be 10× cheaper than GPT-4o without gimmicks.

Every term you'll hit reading the DeepSeek docs or debugging an integration: token, MoE, cache-hit, streaming, tool use, and more.

How deepseek-chat and deepseek-reasoner compare with GPT-4o and o-series on real coding tasks: quality, speed, and cost-per-task.

Why 429s spike, how DeepSeek meters usage, and the retry pattern that keeps a production workload healthy without hammering the API.

What 'reasoning' means in R1, why it streams reasoning_content separately, and how to actually use it in an app.