Side-by-side, with the numbers that matter.

Two rigs, one table. Higher numbers win in cobalt; TDP and price flip.

Rig A

Alibaba · cloud-model

Change

Rig B

NVIDIA · workstation

Change

0 of 38 fields comparable

Spec	Alibaba Qwen2.5-72B-Instruct (Q4 quant)	NVIDIA DGX B200
Bluefield3 Dpus	—	2
Context Window Tokens K	128	—
CPU Base GHZ	—	2.1
CPU Count	—	2
CPU MAX Boost GHZ	—	4
CPU Model	—	Intel Xeon Platinum 8570
Dual Card Count	2	—
Dual Card Vram GB	24	—
FP4 Pflops Dense	—	72
FP4 Pflops Sparse	—	144
FP8 Pflops	—	72
GPT Version	5	—
Gpus	—	8
HBM Bandwidth TB S	—	64
Height MM	—	444
Internal Storage TB	—	30.72
Length MM	—	897.1
MAX Power KW	—	14.3
Networking Osfp Ports	—	4
Nvswitch Count	—	2
OS	—	NVIDIA DGX OS / Ubuntu
OS Storage TB	—	3.8
Parameter Count B	72	—
PER Port Gbps MAX	—	400
Qwen Version	3	—
Qwen25 72B Instruct	Qwen2.5-72B-Instruct	—
Qwen25 Version	2.5	—
Qwen3 Largest Dense Checkpoint B	32	—
Rack Units	—	10
RTX 5090 Pairing Vram GB	32	—
Status Checked	2026-09-24	—
System Memory GB Default	—	2,048
System Memory GB MAX	—	4,096
Total CPU Cores	—	112
Total Hbm3e GB	—	1,440
Total Nvlink Bandwidth TB S	—	14.4
Vram Required GB	44	—
Width MM	—	482.2