DeepSeek-Coder-V2
Code
DeepSeek's code-tuned MoE. Strong on competitive coding benchmarks; the open-weights answer to early Codestral.
DeepSeek-Coder-V2 is a efficient AI model from DeepSeek. It costs $0.140 per million input tokens and $0.280 per million output tokens (blended $0.182/M), with a 128,000-token context window.
INPUT
$0.140/M
per million input tokens
OUTPUT
$0.280/M
per million output tokens
CONTEXT
128,000
tokens
What it is good at
- Code-tuned MoE
- Open weights
- 128K context
Typical use cases
- Self-hosted code completion
- Code review bots
Benchmarks
vs. best public score
Hand-curated from each provider's published reports and public leaderboards. Methodology varies across sources, treat as directional rather than authoritative.
More from DeepSeek
See all 11 →DeepSeek-V4-Pro
Frontier · 1,000,000 ctx
in $1.740/Mout $3.480/M
DeepSeek-V4-Flash
Efficient · 1,000,000 ctx
in $0.140/Mout $0.280/M
DeepSeek-V3
Frontier · 64,000 ctx
in $0.270/Mout $1.100/M
DeepSeek-V3-0324
Frontier · 128,000 ctx
in $0.270/Mout $1.100/M
DeepSeek-R1
Reasoning · 64,000 ctx
in $0.550/Mout $2.190/M
DeepSeek-R1-0528
Reasoning · 64,000 ctx
in $0.550/Mout $2.190/M
Frequently asked questions
How much does DeepSeek-Coder-V2 cost?
DeepSeek-Coder-V2 costs $0.140 per million input tokens and $0.280 per million output tokens, for a blended reference rate of $0.182 per million tokens.
What is DeepSeek-Coder-V2's context window?
DeepSeek-Coder-V2 supports up to 128,000 tokens of context in a single request.
What is DeepSeek-Coder-V2 best for?
DeepSeek-Coder-V2 is well suited to Code-tuned MoE, Open weights and 128K context.
Who makes DeepSeek-Coder-V2?
DeepSeek-Coder-V2 is developed and served by DeepSeek. It was released in Jun 2024.