Skip to content
local-ai

NVIDIA GeForce RTX 4090

NVIDIA · Enthusiast

The previous-generation consumer flagship, now succeeded by the RTX 5090 but still an excellent local AI card, with 24GB of fast GDDR6X memory and strong compute. Enough to run capable models at good quantisations, though 24GB sets a real ceiling on model size.

Specifications

VRAM 24GB
Memory bandwidth 1008 GB/s
Power draw 450W
Type gpu

Prices have moved around a lot with demand. Treat any figure as indicative and check current retail before buying.

NVIDIA GeForce RTX 4090: common questions

What models can the NVIDIA GeForce RTX 4090 run?
With 24GB of VRAM it can run models up to roughly 37B 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 4090 draw?
About 450W under load, so pair it with a power supply that has real headroom.

Strengths

  • Fast memory bandwidth, which matters a great deal for inference speed
  • 24GB comfortably runs strong models up to roughly 32B at good quantisations
  • Mature CUDA support across every major inference engine

Weaknesses

  • 24GB rules out larger models without heavy quantisation or a second card
  • High power draw and heat under sustained load
  • Enthusiast pricing, and availability has been inconsistent
See what models fit in 24GB →

Entry last verified 30 July 2026. Specifications and especially prices change; verify before buying.