DeepSeek • 671B MoE (37B active)128,000 Token Context

DeepSeek R1 (Reasoning) Token Pricing & VRAM Calculator

Breakthrough open-weights reinforcement-learning reasoning model rivaling OpenAI o1. Ideal for: Deep mathematical reasoning, algorithmic logic, open-weights self-hosting.

API Cost (per 1M Tokens)

$0.55 / 1M Input

$2.19 / 1M Output

100k input tokens = ~$0.0550 | 100k output = ~$0.2190

Local VRAM & GPU Specs

40 GB

2x RTX 3090 (Quantized) or 8x H100

Local Ollama Run

ollama run deepseek-r1

Inspect port 11434: Port 11434 Guide

Compare Other AI Foundation Models

GPT-4o (Omni)$2.5/1M in Claude 3.5 Sonnet$3/1M in Llama 3.3 70B Instruct$0.7/1M in Qwen 2.5 Coder 32B$0.3/1M in Mistral Small 3 (24B)$0.2/1M in

Frequently Asked Questions

How much does DeepSeek R1 (Reasoning) cost per 1M tokens?

DeepSeek R1 (Reasoning) costs $0.55 per million input tokens and $2.19 per million output tokens.

What GPU is needed to run DeepSeek R1 (Reasoning) locally?

DeepSeek R1 (Reasoning) requires at least 40GB VRAM for 4-bit quantization. Recommended hardware: 2x RTX 3090 (Quantized) or 8x H100.

What is the context window size of DeepSeek R1 (Reasoning)?

DeepSeek R1 (Reasoning) supports a context window of 128,000 tokens.