DeepSeek logo

DeepSeek Reasoner

deepseek/deepseek-reasoner

DeepSeek reasoning model for multi-step analysis, math, coding, and tools

Input modalitiesT
Output modalitiesT
Source-linkeddeepseek-thinkingOpen weightsReleased 2025-12-01
Report data
Context window1M
Maximum output384K
Provider offers3
Knowledge cutoff2025-09
Capabilities

What this model supports

ReasoningYes
Tool callingYes
Structured outputNo
AttachmentsYes
Temperature controlYes
Open weightsYes
Creator
DeepSeek
Family
deepseek-thinking
Modalities
Text
License
Not documented
Release date
2025-12-01
Model ID
deepseek/deepseek-reasoner
Market availability

Provider pricing

Report price →
Providers offering DeepSeek Reasoner
ProviderProvider model IDContextInput / 1MOutput / 1MCached / 1MFeaturesAll pricing
DeepSeek logoDeepSeekdeepseek-reasoner1M$0.14 if offer.cost.input is defined else '—'}}$0.28 if offer.cost.output is defined else '—'}}$0.003 if offer.cost.cache_read is defined else '—'}}R T ·
View
{
  "cache_read": 0.0028,
  "input": 0.14,
  "output": 0.28
}
OrcaRouter logoOrcaRouterdeepseek/deepseek-reasoner1M$0.435 if offer.cost.input is defined else '—'}}$0.87 if offer.cost.output is defined else '—'}}$0.028 if offer.cost.cache_read is defined else '—'}}R T ·
View
{
  "cache_read": 0.028,
  "input": 0.435,
  "output": 0.87
}
Helicone logoHeliconedeepseek-reasoner128K$0.56 if offer.cost.input is defined else '—'}}$1.68 if offer.cost.output is defined else '—'}}$0.07 if offer.cost.cache_read is defined else '—'}}· · ·
View
{
  "cache_read": 0.07,
  "input": 0.56,
  "output": 1.68
}
Evidence

Published benchmark results

Aider Polyglot vs top 7 models
GPT-5
88.0
o3-pro
84.9
Gemini 2.5 Pro
83.1
o3
81.3
DeepSeek Reasoner
74.2
Claude Opus 4
72.0
o4-mini
72.0