NVIDIA CMP 170HX x2 vs RTX 3090 x2

GPU AGPU B

Best verified output tok/s per model (single-stream generation). Hover a value for the quant.

Verified data on this pair is thin — most of the board is still pending review. Submit your own run and help fill the matrix: one prompt, three steps, no downloads.

NVIDIA CMP 170HX x21 verified run · best 186.9 tok/s

RTX 3090 x26 verified runs · best 240.8 tok/s

Model	NVIDIA CMP 170HX x2	RTX 3090 x2
Qwen3.6-27B	—	240.8 AutoRound INT4
Qwen3.8-27B	—	219.8 AutoRound INT4 + fp8 KV
NVIDIA-Nemotron-3.5-Lightning-30B-A3B	—	197.0 NVFP4
Qwen3.8-Flash-Next	186.9 W4A16-FP8PLE	40.1 Unsloth-Dynamic-IQ3_XXS
Qwen3.8-Flash-Next-REAP-256-duo	—	53.5 Unsloth-Dynamic-Q3_K_XL

Head-to-head (models tested on both): NVIDIA CMP 170HX x2 1 · RTX 3090 x2 0 · tied 0 · data as of Sep 23, 2026. Full rules: methodology.

Related reading: RTX 5090 vs 3090 · quantization · Value index · Full board