Skip to main content
Models/DeepSeek/DeepSeek R1

DeepSeek R1

DeepSeek 🧠 Language Model

V3.2 thinking mode. IMO/IOI/ICPC gold medal performance. Rivals Gemini 3 Pro on reasoning.

⏹ Retired on 2026-07-24

Specifications

Input modalitiestext
Output modalitiestext
Capabilitieschat, reasoning
Relative speed●●●○○ Average
Sunset on2026-07-24
Context window131,072 tokens
Max output65,536 tokens
Vision inputNo
Tool callingNo
Streaming outputYes
Input price$0.245 / 1M
Output price$0.49 / 1M

Pricing

$0.245 / 1M input tokens  ·  $0.49 / 1M output tokens

Platform credit pricing — live rates on the Models page.

More from DeepSeek

D
DeepSeek V4 Pro
DeepSeek flagship V4. 512K context, advanced reasoning + tool use.
🧠 Language Model
D
DeepSeek V4.1 Flash
DeepSeek V4.1 Flash, the fast daily driver that replaced V4 Flash on 2026-09-10. 1M context, thinking (default) and non-thinking modes.
🧠 Language Model
D
DeepSeek V3.2
Daily driver at GPT-5 level. First model to integrate thinking into tool-use. Fast and cost-effective.
🧠 Language Model
D
DeepSeek V4 Flash (legacy name)
Retired upstream on 2026-09-10; DeepSeek serves this name with V4.1 Flash at Flash prices. Use deepseek-flash.
🧠 Language Model