What are the hardware requirements for TinyMistral 248M Chat V3?

TinyMistral 248M Chat V3 requires approximately 0.5 GB of VRAM and supports a context window of 2K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed TinyMistral 248M Chat V3 and how large is it?

TinyMistral 248M Chat V3 is developed by Felladrin, a model with 248m 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 TinyMistral 248M Chat V3?

TinyMistral 248M Chat V3 is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

TinyMistral 248M Chat V3 by Felladrin — VRAM 0.5GB, 2K context

Name: TinyMistral 248M Chat V3
Author: Felladrin

TinyMistral 248M Chat V3 is an open-source language model by Felladrin. Features: 248m LLM, VRAM: 0.5GB, Context: 2K, License: apache-2.0, Instruction-Based, LLM Explorer Score: 0.1.

Autotrain compatible Base model:finetune:locutusque... Base model:locutusque/tinymist... Conversational Dataset:cohereforai/aya datase... Dataset:databricks/databricks-... Dataset:euclaise/reddit-instru... Dataset:felladrin/chatml-aya d... Dataset:felladrin/chatml-capyb... Dataset:felladrin/chatml-datab... Dataset:felladrin/chatml-deita... Dataset:felladrin/chatml-openo... Dataset:felladrin/chatml-reddi... Dataset:felladrin/chatml-ultra... Dataset:hkust-nlp/deita-10k-v0 Dataset:huggingfaceh4/ultracha... Dataset:huggingfaceh4/ultrafee... Dataset:ldjnr/capybara Dataset:open-orca/openorca En Endpoints compatible Instruct Mistral Region:us Safetensors

Model Card on HF 🤗: https://huggingface.co/Felladrin/TinyMistral-248M-Chat-v3

TinyMistral 248M Chat V3 Parameters and Internals

LLM Name	TinyMistral 248M Chat V3
Repository 🤗	https://huggingface.co/Felladrin/TinyMistral-248M-Chat-v3
Base Model(s)	TinyMistral 248M Locutusque/TinyMistral-248M
Model Size	248m
Required VRAM	0.5 GB
Updated	2025-03-31
Maintainer	Felladrin
Model Type	mistral
Instruction-Based	Yes
Model Files	0.5 GB
Supported Languages	en
Model Architecture	MistralForCausalLM
License	apache-2.0
Context Length	2048
Model Max Length	2048
Transformers Version	4.49.0
Tokenizer Class	LlamaTokenizer
Padding Token	[PAD]
Vocabulary Size	32005
Torch Data Type	bfloat16

Rank the TinyMistral 248M Chat V3 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 TinyMistral 248M Chat V3

Best Alternatives	Context / RAM	Downloads	Likes
TinyMistral 248M V2.5 Instruct	32K / 1 GB	2	11
...adrin TinyMistral248M Instruct	32K / 0.5 GB	108	7
TinyMistral 248M Instruct	32K / 1 GB	69	12
Tinymistv1	32K / 0.5 GB	17	0
...istral 248M V2.5 Instruct Orpo	32K / 0.5 GB	6	0
TinyMistral 248M V2 Instruct	32K / 0.5 GB	159	7
...stral V2 Pycoder Instruct 248M	32K / 1 GB	8	3
...mistral 248M Hypnosis Instruct	32K / 0.5 GB	4	2
...al V2 Pycoder Instruct 248M V1	32K / 0.5 GB	11	1
...istral Magicoder Instruct 248M	32K / 0.5 GB	6	2

Note: green Score (e.g. "73.2") means that the model is better than Felladrin/TinyMistral-248M-Chat-v3.