Comparisons

60 GPU · 44 model · 11 app comparisons — VRAM, bandwidth, benchmarks, and compatibility, side by side.

Compare any two GPUs

Pick any two of the 107 GPUs we track for a quick VRAM, bandwidth, and model-compatibility comparison.

GPU A

GPU B

Spec	NVIDIA RTX 4090	NVIDIA RTX 3090
VRAM	24 GB	24 GB
Bandwidth	1008 GB/s	936 GB/s
Backend	CUDA	CUDA
Tier	Consumer	Consumer
Models (native)	44 / 81	44 / 81

Model	NVIDIA RTX 4090	NVIDIA RTX 3090
Llama 3.3 70B Instruct(70B)	—	—
Qwen 3.6 27B(27B)	43.4 t/s(NVFP4)	40.3 t/s(NVFP4)
Llama 3.1 8B Instruct(8B)	38.4 t/s(BF16)	35.6 t/s(BF16)
Qwen 2.5 7B Instruct(7.6B)	41.8 t/s(BF16)	38.8 t/s(BF16)

Estimated tokens/sec at 8k context, assuming the model fits natively in VRAM.

See the full NVIDIA RTX 4090 vs NVIDIA RTX 3090 comparison — analysis, FAQ, and every compatible model →