Skip to content
local-ai

NVIDIA GeForce RTX 4080 Super

NVIDIA · High-end

A high-end Ada Lovelace consumer card with 16GB of memory, a strong all-rounder for local models that fit within 16GB.

Specifications

VRAM 16GB
Memory bandwidth 736 GB/s
Power draw 320W
Type gpu

An upper-tier consumer card. Verify current pricing.

NVIDIA GeForce RTX 4080 Super: common questions

What models can the NVIDIA GeForce RTX 4080 Super run?
With 16GB of VRAM it can run models up to roughly 23B parameters at a 4-bit quantisation, or smaller models with more context. Use the hardware matrix for specifics; these figures are approximate.
How much power does the NVIDIA GeForce RTX 4080 Super draw?
About 320W under load, so pair it with a power supply that has real headroom.

Strengths

  • 736 GB/s bandwidth and fast Ada tensor cores give good speed on models that fit 16GB
  • Broad CUDA support across every major local inference engine
  • Also strong for local image generation and mixed gaming use

Weaknesses

  • 16GB limits it to roughly 13B-class models, or heavy offload for larger
  • A high 320W draw for the memory you get
  • GDDR6X runs hot, and consumer cards lack ECC
See what models fit in 16GB →

Featured in builds

Entry last verified 18 August 2026. Specifications and especially prices change; verify before buying.