What are the hardware requirements for Replit V2 CodeInstruct 3B?

Replit V2 CodeInstruct 3B requires approximately 10.4 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Replit V2 CodeInstruct 3B and how large is it?

Replit V2 CodeInstruct 3B is developed by teknium, a model with 3b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Replit V2 CodeInstruct 3B?

Replit V2 CodeInstruct 3B is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Replit V2 CodeInstruct 3B by teknium — VRAM 10.4GB

Name: Replit V2 CodeInstruct 3B
Author: teknium

Replit V2 CodeInstruct 3B is an open-source language model by teknium. Features: 3b LLM, VRAM: 10.4GB, License: cc-by-sa-4.0, LLM Explorer Score: 0.07.

Code Custom code Dataset:bigcode/the-stack-dedu... Dataset:sahil2801/codealpaca-2... Dataset:teknium/gpteacher-code... Endpoints compatible Instruct Mpt Pytorch Region:us Self-instruct Sharded

Model Card on HF 🤗: https://huggingface.co/teknium/Replit-v2-CodeInstruct-3B

Replit V2 CodeInstruct 3B Parameters and Internals

Model Type

code, instruct, self instruct

Supported Languages

Markdown (proficient), Java (proficient), JavaScript (proficient), Python (proficient), TypeScript (proficient), PHP (proficient), SQL (proficient), JSX (proficient), reStructuredText (proficient), Rust (proficient), C (proficient), CSS (proficient), Go (proficient), C++ (proficient), HTML (proficient), Vue (proficient), Ruby (proficient), Jupyter Notebook (proficient), R (proficient), Shell (proficient)

Training Details

Data Sources:

bigcode/the-stack-dedup, sahil2801/CodeAlpaca-20k, teknium/GPTeacher-CodeInstruct

Data Volume:

~25,000 code instruction/response pairs

Methodology:

fine-tuning

Context Length:

2000

Training Time:

1 hour

Hardware Used:

2x a100 80gb

Input Output

Input Format:

### Instruction: ### Input: ### Response: or ### Instruction: ### Response:

Performance Tips:

Use specified sampler settings and tokenizer decode arguments for optimal performance.

LLM Name	Replit V2 CodeInstruct 3B
Repository 🤗	https://huggingface.co/teknium/Replit-v2-CodeInstruct-3B
Model Size	3b
Required VRAM	10.4 GB
Updated	2026-07-23
Maintainer	teknium
Model Type	mpt
Model Files	10.0 GB: 1-of-2 0.4 GB: 2-of-2
Supported Languages	code
Model Architecture	MPTForCausalLM
License	cc-by-sa-4.0
Model Max Length	2000
Transformers Version	4.29.2
Tokenizer Class	ReplitLMTokenizer
Padding Token	<\|pad\|>
Vocabulary Size	32768
Torch Data Type	float16

Rank the Replit V2 CodeInstruct 3B 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

Quantized Models of the Replit V2 CodeInstruct 3B

Model	Likes	Downloads	VRAM
Replit V2 CodeInstruct 3B GGML	33	2	1 GB

Best Alternatives to Replit V2 CodeInstruct 3B

Best Alternatives	Context / RAM	Downloads	Likes
Replit Code V1.5 3B	0K / 6.6 GB	544	316
Sea Lion 3B	0K / 6.4 GB	778	17
Code Millenials 3B	0K / 5.2 GB	5	1
SEA LION V1 3B	0K / 6.4 GB	186	19
Mpt 3B 8K Instruct	0K / 6.9 GB	5	3
Replit Code V1 3B	0K / 10.4 GB	111	744
Glaive Function Calling V1	0K / 10.4 GB	8	75
...aive Function Calling V2 Small	0K / 10.4 GB	8	15
Evol Replit V1	0K / 10.4 GB	9	8
Replit CodeInstruct V3	0K / 10.4 GB	7	2

Note: green Score (e.g. "73.2") means that the model is better than teknium/Replit-v2-CodeInstruct-3B.