What are the hardware requirements for Codestral 22B V0.1 Hf?

Codestral 22B V0.1 Hf requires approximately 44.7 GB of VRAM and supports a context window of 32K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Codestral 22B V0.1 Hf and how large is it?

Codestral 22B V0.1 Hf is developed by bullerwins, a model with 22b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Codestral 22B V0.1 Hf perform on standard benchmarks?

Codestral 22B V0.1 Hf has the following published scores: MMLU 62.21. Compare against reference models on this page or on the LLM Explorer leaderboards.

Codestral 22B V0.1 Hf by bullerwins — VRAM 44.7GB, 32K context

Name: Codestral 22B V0.1 Hf
Author: bullerwins

Codestral 22B V0.1 Hf is an open-source language model by bullerwins. Features: 22b LLM, VRAM: 44.7GB, Context: 32K, License: MNPL-0.1, LLM Explorer Score: 0.14, Arc: 62.5, HellaSwag: 81.8, MMLU: 62.2, GSM8K: 62.

Code Deploy:azure Mistral Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/bullerwins/Codestral-22B-v0.1-hf

Codestral 22B V0.1 Hf Parameters and Internals

Model Type

code, FIM (Fill in the Middle)

Use Cases

Areas:

software development, code generation, programming assistance

Applications:

VS Code add-ons

Primary Use Cases:

answer questions about code snippets, generate code, document code, factorize code

Limitations:

does not have any moderation mechanisms

Considerations:

Supported Languages

languages_supported (80+ programming languages, including Python, Java, C, C++, JavaScript, and Bash.), proficiency_levels ()

Release Notes

Version:

0.1

Date:

31/05/2024

Notes:

Fixed the tokenizer issue.

LLM Name	Codestral 22B V0.1 Hf
Repository 🤗	https://huggingface.co/bullerwins/Codestral-22B-v0.1-hf
Model Size	22b
Required VRAM	44.7 GB
Updated	2026-07-17
Maintainer	bullerwins
Model Type	mistral
Model Files	4.9 GB: 1-of-9 5.0 GB: 2-of-9 5.0 GB: 3-of-9 4.9 GB: 4-of-9 5.0 GB: 5-of-9 5.0 GB: 6-of-9 4.9 GB: 7-of-9 5.0 GB: 8-of-9 5.0 GB: 9-of-9
Supported Languages	code
Model Architecture	MistralForCausalLM
License	MNPL-0.1
Context Length	32768
Model Max Length	32768
Transformers Version	4.40.2
Tokenizer Class	LlamaTokenizer
Vocabulary Size	32768
Torch Data Type	bfloat16

Rank the Codestral 22B V0.1 Hf 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 Codestral 22B V0.1 Hf

Model	Likes	Downloads	VRAM
Codestral 22B V0.1 AWQ	0	6876	12 GB
Codestral 22B V0.1 GGUF	9	2720	4 GB
Codestral 22B V0.1 Hf AWQ	2	402	12 GB
Codestral 22B V0.1 AWQ	0	113	12 GB
Codestral 22B V0.1 GPTQ	0	23	12 GB
Codestral 22B V0.1 Hf AWQ	0	7	12 GB

Best Alternatives to Codestral 22B V0.1 Hf

Best Alternatives	Context / RAM	Downloads	Likes
MS Schisandra 22B V0.2	128K / 44.7 GB	8	9
...ntheon RP Pure 1.6.2 22B Small	128K / 44.7 GB	17	35
MS Meadowlark 22B	128K / 44.7 GB	13	16
...rker The Final Abomination 22B	128K / 44.7 GB	6	6
...er The Final Transgression 22B	128K / 44.7 GB	3	3
...Darker The Final Directive 22B	128K / 44.7 GB	6	0
Retrograde Omega M 22B V1.0	128K / 44.7 GB	5	0
The Omega Directive M 22B V1.0	128K / 44.7 GB	12	2
MS Inky 2409 22B	128K / 44.7 GB	5	0
MS Springydragon 2409 22B	128K / 44.7 GB	8	0

Note: green Score (e.g. "73.2") means that the model is better than bullerwins/Codestral-22B-v0.1-hf.