DeepSeek: R1 — pricing & benchmarks

By deepseek · Updated 2026-10-06 · Open weights · mitstep-by-step reasoningopen weights / self-hostable

What DeepSeek: R1 is used for

DeepSeek: R1 is best suited for step-by-step reasoning, open weights / self-hostable. Access it through one Appmo API key alongside 458 other models — Appmo's optimizer can route to it automatically when it is the best value for a request. Weights: huggingface.co/deepseek-ai/DeepSeek-R1.

Use it via the Appmo API (OpenAI-compatible)

curl https://appmo.com/v1/chat/completions \
  -H "Authorization: Bearer $APPMO_KEY" \
  -d '{"model":"deepseek/deepseek-r1","messages":[{"role":"user","content":"Hello"}]}'

FAQ

How much does DeepSeek: R1 cost?
DeepSeek: R1 costs $0.7 per 1M input tokens and $2.5 per 1M output tokens ($1.15 blended), as of 2026-10-06. Prices are expected to fall as competition increases.

What is DeepSeek: R1 best for?
DeepSeek: R1 is best suited for step-by-step reasoning, open weights / self-hostable.

How do I use DeepSeek: R1?
Send requests to Appmo's OpenAI-compatible endpoint with model id "deepseek/deepseek-r1" and one API key — no separate provider account needed.

Is DeepSeek: R1 open source?
Yes — its weights are public on Hugging Face (deepseek-ai/DeepSeek-R1) under the mit license; check the license terms before commercial use.

Can DeepSeek: R1 create images?
No, DeepSeek: R1 works with text.