Best GPU for local LLMs under $2,000
RTX 4090, RTX 5080, used 3090, RX 7900 XTX, Arc B580. Here's where each lands on real workloads.
No under-$2,000 GPU submissions yet. The tier covers the RTX 5090 at MSRP, the used 4090, dual 3090s, the AMD RX 7900 XTX, and the new 5070 Ti — be the first to submit one and this section will lead with the headline number from your run.
No data submitted for this task yet.
Run the suite to be the first benchmark for this guide:
$ pipx install llm-speed && llm-speed bench
The under-$2k tier is the sweet spot for prosumer local LLMs. You're choosing among a new RTX 5090 at MSRP, a used 4090, a pair of 3090s, an AMD RX 7900 XTX, or a budget Arc B580. Each has a different memory-bandwidth-to-VRAM ratio and a different software story (CUDA vs ROCm vs SYCL). Below are submitted benchmarks across this tier.
Side-by-side comparisons
See also: All hardware · All models · Methodology