D

DeepSeek

Major Foundation Lab22 Canonical Models

DeepSeek foundation models, inference API endpoints, and pricing.

Official Provider Docs
Endpoint: https://api.deepseek.com•Package: @ai-sdk/openai-compatible•Env Key: DEEPSEEK_API_KEY

Models by DeepSeek

Specifications, context windows, and pricing for DeepSeek models

22 models
DeepSeek V4.1 Flash 1017
deepseek/deepseek-v4.1-flash-1017•DeepSeek
1M ctx

October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding

384K out maxVisionHumanity's Last Exam: 42.6 (unit not reported)
In: $0.15 • Out: $0.60
DeepSeek
DeepSeek V4 Pro 0813
deepseek/deepseek-v4-pro-0813•DeepSeek
1M ctx

DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes

384K out maxHumanity's Last Exam: 42.7 score
In: $2.00 • Out: $4.00
DeepSeek
DeepSeek V4 Pro 0423
deepseek/deepseek-v4-pro-0423•DeepSeek
1M ctx

DeepSeek V4 Pro initial snapshot with million-token context and support for thinking and non-thinking modes

384K out max
In: $1.32 • Out: $3.96
DeepSeek
DeepSeek V4 Pro
deepseek/deepseek-v4-pro•DeepSeek
1M ctx

Open MoE flagship with million-token context for coding and long agent runs

384K out maxSWE-Bench Verified: 80.6 resolved
In: $0.66 • Out: $1.98
DeepSeek
DeepSeek V4 Flash 0423
deepseek/deepseek-v4-flash-0423•DeepSeek
1M ctx

Initial DeepSeek V4 Flash snapshot for economical reasoning, coding, and million-token agent workloads

384K out max
In: $0.14 • Out: $0.28
DeepSeek
DeepSeek V4 Flash 0731
deepseek/deepseek-v4-flash-0731•DeepSeek
1M ctx

Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding

384K out maxTerminal-Bench: 82.7 pass@1
In: $0.08 • Out: $0.41
DeepSeek
DeepSeek V4 Flash Vision Exp
deepseek/deepseek-v4-flash-vision-exp•DeepSeek
1M ctx

Experimental multimodal DeepSeek V4 Flash model for image understanding, coding, and agentic work

384K out maxVisionTerminal-Bench: 83.9 score
In: $0.15 • Out: $0.60
DeepSeek
DeepSeek Chat
deepseek/deepseek-chat•DeepSeek
1M ctx

DeepSeek chat model for instruction following, coding, and analysis

384K out maxAider Polyglot: 70.2 percent correct
In: $0.10 • Out: $0.42
DeepSeek
DeepSeek V4.1 Flash
deepseek/deepseek-v4.1-flash•DeepSeek
1M ctx

DeepSeek V4.1 Flash model for reasoning and agentic coding

384K out maxVisionGPQA Diamond: 90.9 pass@1
In: $0.51 • Out: $1.53
DeepSeek
DeepSeek V4 Flash
deepseek/deepseek-v4-flash•DeepSeek
1M ctx

Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work

384K out maxSWE-Bench Verified: 79 resolved
In: $0.15 • Out: $0.60
DeepSeek
DeepSeek Reasoner
deepseek/deepseek-reasoner•DeepSeek
1M ctx

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

384K out maxAider Polyglot: 74.2 percent correct
In: $0.56 • Out: $1.68
DeepSeek
DeepSeek R1 0528
deepseek/deepseek-r1-0528•DeepSeek
163.8K ctx

May 2025 DeepSeek R1 update with improved mathematics, coding, and reasoning

32.8K out max
In: $0.50 • Out: $2.15
DeepSeek
DeepSeek V3 0324
deepseek/deepseek-v3-0324•DeepSeek
163.8K ctx

March 2025 checkpoint of DeepSeek-V3 with improved reasoning and coding

163.8K out max
In: $0.20 • Out: $0.77
DeepSeek
DeepSeek V3 0324
deepseek/deepseek-chat-v3-0324•DeepSeek
163.8K ctx

DeepSeek chat model for instruction following, coding, and analysis

147.5K out max
In: $0.28 • Out: $1.10
DeepSeek
DeepSeek V3.1
deepseek/deepseek-chat-v3.1•DeepSeek
163.8K ctx

DeepSeek chat model for instruction following, coding, and analysis

32.8K out max
In: $0.20 • Out: $0.70
DeepSeek
DeepSeek-V3
deepseek/deepseek-v3•DeepSeek
131.1K ctx

Open DeepSeek MoE chat model for coding, math, and general reasoning

8.2K out max
In: $0.50 • Out: $1.50
DeepSeek
DeepSeek R1 0528 Qwen3 8B
deepseek/deepseek-r1-0528-qwen3-8b•DeepSeek
131.1K ctx

DeepSeek R1 0528 distilled into Qwen3 8B for mathematics, coding, and reasoning

32K out max
In: $0.06 • Out: $0.09
DeepSeek
DeepSeek-V3.1
deepseek/deepseek-v3.1•DeepSeek
131.1K ctx

Hybrid-reasoning DeepSeek model with thinking and non-thinking modes

8.2K out max
In: $0.19 • Out: $0.71
DeepSeek
DeepSeek-R1-Distill-Qwen-32B
deepseek/deepseek-r1-distill-qwen-32b•DeepSeek
131.1K ctx

R1 reasoning distilled into Qwen 2.5 32B for efficient open-weight step-by-step problem solving

32.8K out max
In: $0.29 • Out: $0.86
DeepSeek
DeepSeek-R1
deepseek/deepseek-r1•DeepSeek
128K ctx

Classic open reasoning model for transparent math, coding, and deliberate problem solving

32.8K out maxAider Polyglot: 56.9 percent correct
In: $0.70 • Out: $2.50
DeepSeek
DeepSeek V3.2
deepseek/deepseek-v3.2•DeepSeek
128K ctx

Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use

64K out max
In: $0.50 • Out: $1.50
DeepSeek
DeepSeek OCR 2
deepseek/deepseek-ocr-2•DeepSeek
8.2K ctx

High-accuracy OCR model for extracting text from documents, screenshots, receipts, and natural scenes

8.2K out maxVision
In: $0.03 • Out: $0.03
DeepSeek