models/meta-llama-llama-3.3-70b-instruct

Llama 3.3 70B

Sources arena · Confidence C

params

70B

dense

bench sources

single

context

—

max tokens

license

Llama-3.3

community

▸Model meta

▸Benchmark breakdown

▸Hardware fit matrix

by gpu tier	Q2_K 26.3 GB	Q4_K_M 35.0 GB	Q5_K_M 43.8 GB	Q6_K 52.5 GB	Q8_0 70.0 GB
8 GB tier rtx 3050 · 4060
12 GB tier rtx 3060 · 4070
16 GB tier rtx 4080 · 4060 ti 16g
24 GB tier rtx 3090 · 4090
32 GB tier rtx 5090 · m3 max	OK 64 t/s
48 GB tier a6000 · m3 max 64	GREAT 21 t/s	GREAT 16 t/s	TIGHT 12 t/s
80 GB tier h100 · m3 ultra 128	GREAT 70 t/s	GREAT 53 t/s	GREAT 42 t/s	GREAT 35 t/s	OK 27 t/s

full fit · production speed tight fit · usabledoesn't fit

Get personalized ranking

Tell us your machine — we'll tell you if this is actually your best pick, or what's better.