All models
DeepSeek

DeepSeek R1

frontierOpen Source

DeepSeek

DeepSeek's open-source reasoning model. Matches o1-level math and coding with full transparency — the thinking chain is visible in the output.

Official page
open-sourcereasoningmath
Context
64K
Input / 1M
$0.70
Output / 1M
$2.50
Knowledge cutoff
2024-07

Capabilities

Code5/5
Reasoning5/5
Multilingual3/5
Safety2/5
Speed2/5

Modalities & Features

Vision (image input) No
Audio No
Video No
Function / tool calling No
JSON mode Yes
Extended thinking Yes

Context & Output

Context window64K
Max output tokens32.768K
Knowledge cutoff2024-07
Release date2025-01

API & Access

Open source Yes
Streaming Yes
Batch API No
Prompt caching Yes
Fine-tuning No

Benchmarks

MMLUGeneral knowledge
90.8%
GPQA DiamondHard science Q&A
71.5%
MATHMath problem solving
83.5%
AIME 2025Advanced math
N/A
GSM8KGrade-school math
97.3%
HumanEvalCode generation
85.7%
LiveCodeBenchLive coding problems
N/A
SWE-bench VerifiedReal-world software
49.2%
HellaSwagCommonsense inference
N/A

Source: provider technical reports & independent evaluations. N/A = not yet published for this model.

Pricing (per 1M tokens)

Input$0.70
Output$2.50
Consumer plan
DeepSeek Chat
Free
Prices shown are list prices and may vary by provider plan. Always check official documentation for the latest pricing.