NelsaHost
search Login
smart_toy

GPUs for Llama 3.3 70B (70.00B)

70.00B parameters • 140.0GB VRAM (FP16) • Llama

info Model Details
check_circle 0 Perfect
thumb_up 0 Good
warning 0 Marginal
cancel 35 Fail
cancel

RTX 3090

NVIDIA • 24.0GB

FAIL/OOM
24.0GB available -116.0GB headroom
cancel

RTX 4090

NVIDIA • 24.0GB

FAIL/OOM
24.0GB available -116.0GB headroom
cancel

RTX 5000 Ada

NVIDIA • 32.0GB

FAIL/OOM
32.0GB available -108.0GB headroom
cancel

Jetson AGX Orin 32GB

NVIDIA • 32.0GB

FAIL/OOM
32.0GB available -108.0GB headroom
cancel

A100 40GB

NVIDIA • 40.0GB

FAIL/OOM
40.0GB available -100.0GB headroom
cancel

RTX A6000

NVIDIA • 48.0GB

FAIL/OOM
48.0GB available -92.0GB headroom
cancel

RTX 6000 Ada

NVIDIA • 48.0GB

FAIL/OOM
48.0GB available -92.0GB headroom
cancel

Jetson AGX Orin 64GB

NVIDIA • 64.0GB

FAIL/OOM
64.0GB available -76.0GB headroom
cancel

A100 80GB

NVIDIA • 80.0GB

FAIL/OOM
80.0GB available -60.0GB headroom
cancel

H100 SXM

NVIDIA • 80.0GB

FAIL/OOM
80.0GB available -60.0GB headroom
cancel

H100 PCIe

NVIDIA • 80.0GB

FAIL/OOM
80.0GB available -60.0GB headroom
Previous Page 2 of 2