How much VRAM does madlad400-3b-mt need?

Q4_K_M is exactly 1,858,124,864 bytes (1.73 GiB) in weights. Add the KV cache, which depends on context length, plus roughly half a gigabyte of runtime overhead.

google · text

madlad400-3b-mt

google/madlad400-3b-mt

madlad400-3b-mt at Q4_K_M is exactly 1,858,124,864 bytes (1.73 GiB / 1.86 GB) — an effective 5.056 bits per weight, not the nominal 4.

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

Parameters

2.9B

Architecture

null layers

Context

—

native (config.json)

License

apache-2.0

Shipped quantizations

● exact bytes, summed from published files

Quant	Size●	Exact bytes●	Effective bpw●	Tensors●	Publisher
Q4_K_M	1.73 GiB	1,858,124,864	5.056	—	notjjustnumbers
Q4_K_M	1.73 GiB	1,858,124,864	5.056	—	mtsdurica
Q4_K	1.90 GiB	2,036,588,352	5.541	—	cstr
Q8_0	3.14 GiB	3,375,619,904	9.184	—	cstr
F16	5.48 GiB	5,886,304,000	16.015	—	cstr

Compare with

same modality, comparable size

vs gemma-4-E4B-it vs Qwen3-4B vs Qwen3-8B vs HyperCLOVAX-SEED-Text-Instruct-1.5B vs Llama-3.2-1B-Instruct vs Qwen3-VL-8B-Instruct-abliterated-v1 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 1.54 GiB. The real file is 1.73 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

—

Vocab

256,000

Sliding window

none

SWA period

—

MLA

Experts

—

Experts per token

—

use_sliding_window

—

Questions people ask

How much VRAM does madlad400-3b-mt need?: Q4_K_M is exactly 1,858,124,864 bytes (1.73 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.
Which quantization of madlad400-3b-mt 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.