Llama-3.1-8B-Instruct vs granite-4.0-h-tiny

Specs, VRAM requirements and download trends — updated 6 August 2026.

meta-llama · A
Params
8.0 B
Context
30d
8.0 M
llama3.1 licence · conditional
ibm-granite · B
Params
6.9 B
Context
131,072
30d
292 K
apache-2.0 · commercial OK

Specification comparison

Differences are highlighted; identical values are muted.

Specification Llama-3.1-8B-Instruct granite-4.0-h-tiny
Parameters 8.0 B 6.9 B
Architecture LlamaForCausalLM GraniteMoeHybridForCausalLM
Context length 131,072
Licence llama3.1 apache-2.0
Commercial use Conditional Allowed
Languages 8
Downloads 30d 8,014,793 291,998
Downloads all time 175.3 M 961 K
Quantizations on the Hub SAFETENSORS SAFETENSORS
Gated Yes — accept licence No
First seen on the Hub 2024-07-18 2025-09-16

Download trend

Daily snapshots, last 10 days (28 Jul – 6 Aug)

8.0 M278 K
Llama-3.1-8B-Instruct granite-4.0-h-tiny

VRAM side by side

On RTX 4090 · 24 GB · 8K context unless noted

Llama-3.1-8B-Instruct · fp16 19.4 GB / 24 GB
✅ Runs comfortably
granite-4.0-h-tiny · fp16 @ 128K ctx 32.4 GB / 24 GB
❌ Won’t fit

Adjacent comparisons