DeepSeek Coder V2
deepseek/deepseek-coder-v2Open coding specialist with fine-tuning support, a cost-effective base for code assistants and CI automation.
128K ctx · Price on request
Category
Coding
Code generation, review and agentic engineering. From autocomplete to autonomous engineering: models tuned for repositories, diffs and tool loops. Reach for a coding model when the work touches a repository rather than a paragraph. Agent-grade use leans hardest on , and is what keeps re-sent context affordable.
4 models in Coding from 4 providers
R/1M tokens These models bill on input alone. Nothing is generated in return, so the output price on each card reads “—”.
Top models
Leading coding models
Browse all coding models →
4 models in the directory
How to choose
Choosing a model in Coding
- 01
Agentic vs completion
Agent-grade coding needs strong tool calling and long context. Simple completion tasks can use far cheaper models.
- 02
Benchmark on your stack
Coding scores vary sharply by language and framework. A model great at Python may lag on your Rust codebase.
- 03
Mind the context economics
Repository-scale prompts are token-heavy. Caching support can cut effective input cost dramatically.