What are the hardware requirements for Code 13B?

Code 13B requires approximately 26.1 GB of VRAM and supports a context window of 4K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Code 13B and how large is it?

Code 13B is developed by ajibawa-2023, a model with 13b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Code 13B perform on standard benchmarks?

Code 13B has the following published scores: MMLU 53.17. Compare against reference models on this page or on the LLM Explorer leaderboards.

Code 13B by ajibawa-2023 — VRAM 26.1GB, 4K context

Name: Code 13B
Author: ajibawa-2023

Code 13B is an open-source language model by ajibawa-2023. Features: 13b LLM, VRAM: 26.1GB, Context: 4K, License: cc-by-nc-nd-4.0, HF Score: 54.8, LLM Explorer Score: 0.14, Arc: 57.3, HellaSwag: 83.3, MMLU: 53.2, TruthfulQA: 42.5, WinoGrande: 73.6, GSM8K: 19.

Code Dataset:ajibawa-2023/code-74k-... En Endpoints compatible Llama Model-index Pytorch Region:us Sharded

Model Card on HF 🤗: https://huggingface.co/ajibawa-2023/Code-13B

Code 13B Benchmarks

ARC: 57.34 vs 96.7 (so35)^-40.7%

HellaSwag: 83.28 vs 95.3 (gpt4)^-12.6%

MMLU: 53.17 vs 88.3 (so35)^-39.8%

TruthfulQA: 42.46 vs 59 (gpt4)^-28%

WinoGrande: 73.56 vs 87.5 (gpt4)^-15.9%

GSM8K: 19.03 vs 96.4 (so35)^-80.3%

LLME Score: 0.14114

^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)

🌟 Advertise your project 🚀

Code 13B Parameters and Internals

Model Type

text-generation

Additional Notes

Extremely thankful to TheBloke for making quantized versions of the model available as GPTQ, GGUF, and AWQ.

Supported Languages

Python (proficient), Java (proficient), JavaScript (proficient), GO (proficient), C++ (proficient), Rust (proficient)

Training Details

Data Sources:

Code-74k-ShareGPT, Python-Code-23k-ShareGPT

Data Volume:

74,000 sets of codes, each with 2 conversations

Methodology:

Full fine-tuning on Llama-2 using DeepSpeed codebase

Training Time:

42 hours

Hardware Used:

Azure 4 x A100 80GB

Model Architecture:

Using Meta's Llama-2

Input Output

Input Format:

Conversation format in Vicuna/ShareGPT v1.1

Accepted Modalities:

text

Output Format:

Text generation with code explanations

Performance Tips:

Prompts can be tailored for specific code languages or tasks.

LLM Name	Code 13B
Repository 🤗	https://huggingface.co/ajibawa-2023/Code-13B
Model Size	13b
Required VRAM	26.1 GB
Updated	2026-03-30
Maintainer	ajibawa-2023
Model Type	llama
Model Files	8.6 GB: 1-of-4 8.5 GB: 2-of-4 8.5 GB: 3-of-4 0.5 GB: 4-of-4
Supported Languages	en
Model Architecture	LlamaForCausalLM
License	cc-by-nc-nd-4.0
Context Length	4096
Model Max Length	4096
Transformers Version	4.28.1
Tokenizer Class	LlamaTokenizer
Beginning of Sentence Token	<s>
End of Sentence Token	</s>
Unk Token	<unk>
Vocabulary Size	32000
Torch Data Type	bfloat16

Quantized Models of the Code 13B

Model	Likes	Downloads	VRAM
Code 13B GGUF	7	234	5 GB
Code 13B GPTQ	1	21	7 GB
Code 13B AWQ	1	16	7 GB

Best Alternatives to Code 13B

Best Alternatives	Context / RAM	Downloads	Likes
Luminaura RP 13B	128K / 26 GB	7	0
Yarn Llama 2 13B 128K	128K / 26 GB	1146	112
Agent Llama2 13B 80K	80K / 26.4 GB	5	0
Chat Llama2 13B 80K	80K / 52.8 GB	11	0
LongAlign 13B 64K	64K / 26 GB	14	13
LongAlign 13B 64K	64K / 26 GB	11	13
LongAlign 13B 64K Base	64K / 26 GB	24	3
LongAlign 13B 64K Base	64K / 26 GB	6	3
Openbuddy Llama2 13B V15p1 64K	64K / 26.1 GB	6	4
Openbuddy Llama2 13b64k V15	64K / 26.1 GB	5	2

Rank the Code 13B Capabilities

🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟

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

What open-source LLMs or SLMs are you in search of? 53310 in total.

Email us: info@extractum.io. Our Privacy Policy | Terms and Conditions | Suggest an improvement.

Our Social Media →

Original data from HuggingFace, OpenCompass and various public git repos.

Check out Ag3ntum — our secure, self-hosted AI agent for server management.

Release v20260328a

Support LLM Explorer

Code 13B by ajibawa-2023