DeepSeek-Coder-V2
ReleasedGPT-4 Turbo level code model with 338 language support
Released on 2024.06.17
Overview
DeepSeek-Coder-V2 is an open-source code model that achieves performance comparable to GPT-4 Turbo in code-specific tasks. It supports 338 programming languages and extends context length to 128K.
Key Features
- GPT-4 Turbo level performance
- 338 programming languages
- 128K context window
- Based on DeepSeek-V2 architecture
Specifications
- Parameters
- 236B (21B activated)
- Architecture
- MoE + MLA
- Context Length
- 128K tokens
- License
- DeepSeek License