What are the hardware requirements for DeepSeek Coder V2 Instruct?

DeepSeek Coder V2 Instruct requires approximately 387 GB of VRAM and supports a context window of 160K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed DeepSeek Coder V2 Instruct and how large is it?

DeepSeek Coder V2 Instruct is developed by deepseek-ai, a model with 235.7b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does DeepSeek Coder V2 Instruct perform on standard benchmarks?

DeepSeek Coder V2 Instruct has the following published scores: LMSYS ELO 1193. Compare against reference models on this page or on the LLM Explorer leaderboards.

DeepSeek Coder V2 Instruct by deepseek-ai — VRAM 387GB, 160K context

Name: DeepSeek Coder V2 Instruct
Rating: 3.67 (3 reviews)
Author: deepseek-ai

DeepSeek Coder V2 Instruct is an open-source language model by deepseek-ai. Features: 235.7b LLM, VRAM: 387GB, Context: 160K, License: other, Instruction-Based, Code Generating, LLM Explorer Score: 0.3, ELO: 1193.

Arxiv:2401.06066 Base model:deepseek-ai/deepsee... Base model:finetune:deepseek-a... Codegen Conversational Custom code Deepseek v2 Endpoints compatible Eval-results Instruct Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/deepseek-ai/DeepSeek-Coder-V2-Instruct

DeepSeek Coder V2 Instruct Benchmarks

LMSys ELO 1193 vs 1272 (so35) -6.2%

^nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

What is the LLM Explorer Rank (Score)

DeepSeek Coder V2 Instruct Parameters and Internals

Model Type

Mixture-of-Experts, code language model

Use Cases

Areas:

code intelligence, mathematical reasoning

Applications:

OpenAI-Compatible API, DeepSeek Platform

Supported Languages

programming_languages_supported (86-338)

Training Details

Data Volume:

6 trillion tokens

Methodology:

Continued pre-training

Context Length:

128000

Model Architecture:

DeepSeekMoE framework

Input Output

Input Format:

Code and chat prompts

Accepted Modalities:

text

Output Format:

Generated codes or chat responses

Release Notes

Version:

Notes:

Expanded support for programming languages from 86 to 338, context length from 16K to 128K, enhanced coding and mathematical reasoning capabilities.

LLM Name	DeepSeek Coder V2 Instruct
Repository 🤗	https://huggingface.co/deepseek-ai/DeepSeek-Coder-V2-Instruct
Base Model(s)	DeepSeek Coder V2 Base deepseek-ai/DeepSeek-Coder-V2-Base
Model Size	235.7b
Required VRAM	387 GB
Updated	2026-07-05
Maintainer	deepseek-ai
Model Type	deepseek_v2
Instruction-Based	Yes
Model Files	8.6 GB: 1-of-55 8.6 GB: 2-of-55 8.6 GB: 3-of-55 8.6 GB: 4-of-55 8.6 GB: 5-of-55 8.6 GB: 6-of-55 8.6 GB: 7-of-55 8.6 GB: 8-of-55 8.6 GB: 9-of-55 8.6 GB: 10-of-55 8.6 GB: 11-of-55 8.6 GB: 12-of-55 8.6 GB: 13-of-55 8.6 GB: 14-of-55 8.6 GB: 15-of-55 8.6 GB: 16-of-55 8.6 GB: 17-of-55 8.6 GB: 18-of-55 8.6 GB: 19-of-55 8.6 GB: 20-of-55 8.6 GB: 21-of-55 8.6 GB: 22-of-55 8.6 GB: 23-of-55 8.6 GB: 24-of-55 8.6 GB: 25-of-55 8.6 GB: 26-of-55 8.6 GB: 27-of-55 8.6 GB: 28-of-55 8.6 GB: 29-of-55 8.6 GB: 30-of-55 8.6 GB: 31-of-55 8.6 GB: 32-of-55 8.6 GB: 33-of-55 8.6 GB: 34-of-55 8.6 GB: 35-of-55 8.6 GB: 36-of-55 8.6 GB: 37-of-55 8.6 GB: 38-of-55 8.6 GB: 39-of-55 8.6 GB: 40-of-55 8.6 GB: 41-of-55 8.6 GB: 42-of-55 8.6 GB: 43-of-55 8.6 GB: 44-of-55 8.6 GB: 45-of-55
Generates Code	Yes
Model Architecture	DeepseekV2ForCausalLM
License	other
Context Length	163840
Model Max Length	163840
Transformers Version	4.39.3
Vocabulary Size	102400
Torch Data Type	bfloat16

Rank the DeepSeek Coder V2 Instruct capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to DeepSeek Coder V2 Instruct

Best Alternatives	Context / RAM	Downloads	Likes
...eepSeek Coder V2 Instruct 0724	160K / 378.4 GB	358	117