HomeModel Memory Fit › Mistral 7B Instruct v0.3
Model Memory Fit · LLM

Mistral 7B Instruct v0.3: memory fit on every Jetson module.

Method v1.0 · dataset 2026-09-07 · verified 2026-09-07 · last updated September 2026

7.25 B parameters · 32 layers · 32,768-token max context. Hugging Face: mistralai/Mistral-7B-Instruct-v0.3.

Model facts

FactValueNote
Parameters7.248 B
Layers32
Hidden size4096
Attention heads32
KV heads8
Head dimension128
Max context32,768 tokens
dtypebfloat16
Vocabulary32,768

Mistral AI · hf:mistralai/Mistral-7B-Instruct-v0.3 · config.json · verified 2026-09-07

“"num_attention_heads": 32, "num_key_value_heads": 8, "vocab_size": 32768”

Parameters source

  • Hugging Face · hf:mistralai/Mistral-7B-Instruct-v0.3 · verified 2026-09-07 · model.safetensors.index.json · class A
    “"total_size": 14496047104”

Published quantised artefacts

A file the publisher or a community mirror actually ships, used as the class-A weight figure when the requested quant matches. Any quant without a row here falls back to parameters × bytes-per-parameter (class D for block quants).

QuantFormatFileSizeSource
GGUF Q4_K_M / INT4 (AWQ, MLC q4f16)GGUF Q4_K_MMistral-7B-Instruct-v0.3-Q4_K_M.gguf4.373 GBbartowski (community GGUF) · hf:bartowski/Mistral-7B-Instruct-v0.3-GGUF · verified 2026-09-07

Jetson tokens/s measurements

Published or archived throughput numbers, not modelled. Class C (external measured benchmark).

No Jetson-measured tokens/s reported for this model yet.

Every module × every quantisation

Verdict and total memory at 4096-token context, 1 sequence, llama.cpp, headless. Q4 cells link to the static breakdown page; every other cell links to the live engine at that quantisation.

Notes

head_dim derived as hidden_size/num_attention_heads=4096/32=128 (no explicit head_dim field in config.json). params_b (7.25B) is higher than the nominal '7B' name due to the extended 32768-token vocabulary vs Mistral-7B-v0.1's 32000.

Method and data: Model Memory Fit methodology. Full registry: model-memory.json.

Change the context, concurrency or runtime.

The live engine covers any context length, sequence count, KV precision and runtime, with a permanent link.

OPEN MODEL MEMORY FIT →