How much VRAM does TinyMistral-248M-v2-Instruct need?

Q4_K_M is exactly 155,668,768 bytes (0.14 GiB) in weights. Add the KV cache, which depends on context length, plus roughly half a gigabyte of runtime overhead.

Locutusque · text

TinyMistral-248M-v2-Instruct

Locutusque/TinyMistral-248M-v2-Instruct

TinyMistral-248M-v2-Instruct at Q4_K_M is exactly 155,668,768 bytes (0.14 GiB / 0.16 GB) — an effective 5.021 bits per weight, not the nominal 4.

From the file· summed from 1 file(s)From the file· KV per layer

Parameters

248M

Architecture

llama

12 layers

Context

32,768

native (config.json)

License

apache-2.0

Shipped quantizations

● exact bytes, summed from published files

Quant	Size●	Exact bytes●	Effective bpw●	Tensors●	Publisher
Q2_K	0.10 GiB	105,458,816	3.402	—	M4-ai
Q3_K_M	0.12 GiB	129,029,760	4.162	—	M4-ai
Q4_K_M	0.14 GiB	155,668,768	5.021	—	M4-ai
Q5_K_M	0.17 GiB	179,229,600	5.781	—	M4-ai
Q6_K	0.19 GiB	204,263,008	6.588	—	M4-ai
Q8_0	0.25 GiB	264,325,088	8.526	—	M4-ai

KV cache by context

unresolved

This model declares a 32-token sliding window, but we could not establish which layers use it. Its architecture publishes the layout as a per-layer array inside the model file rather than as a period in config.json, and we have not yet ingested that array.

A flat context × layers × heads figure would be substantially too high, so we are not showing one. This is tracked as a known gap rather than filled with a guess.

Compare with

same modality, comparable size

vs ced-base vs Qwen3-0.6B vs SmolLM2-135M-Instruct vs gemma-3-270m-it vs tinygemma3_cifar vs LFM2.5-230M All alternatives →Will it fit on my card? →

Why other calculators give a different number

A parameters × bits ÷ 8 estimate puts Q4_K_M at roughly 0.13 GiB. The real file is 0.14 GiB, because a quantization is a mixture and some tensors are always kept at higher precision.

Architecture

from config.json

Layers

Attention heads

KV heads

Head dim

Hidden size

1024

Vocab

32,005

Sliding window

SWA period

—

MLA

Experts

—

Experts per token

—

use_sliding_window

—

Questions people ask

How much VRAM does TinyMistral-248M-v2-Instruct need?: Q4_K_M is exactly 155,668,768 bytes (0.14 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.
Which quantization of TinyMistral-248M-v2-Instruct should I use?: Q4_K_M is the usual default. Pick the largest quantization that fits your card at the context you actually need — the table above gives exact sizes for every one published.