RTX PRO 6000 server

Air-cooled GPU server, Blackwell Server Edition

Short answer. The RTX PRO 6000 server carries the Blackwell Server Edition GPU, air-cooled in a standard chassis. It serves several models of 70 to 200 billion parameters in parallel, RAG and multi-GPU fine-tuning, with vLLM or the ninfer runtime. Its cost per token remains to be measured on the target model (see the price guide).

Hardware sheet · QDNA sovereign AI platform · Published on 30 July 2026 · Updated on 13 October 2026

Indicative price	on quotation: no manufacturer publishes a price for an 8-card server; only the card itself is listed, $16,000 excl. VAT on the NVIDIA marketplace as of 1 September 2026 according to thundercompute.com (price guide)
Memory	96 GB GDDR7 per GPU (2 to 8 GPUs, up to 768 GB)
Compute	4 PFLOPS FP4 with sparsity per GPU (2 PFLOPS FP8)
Bandwidth	1,597 GB/s per card, PCIe Gen5
Isolation	MIG up to four 24 GB instances
Power draw	up to 600 W per GPU (configurable), air-cooled
Density	2 to 8 GPUs per server

What the machine runs

Several models from 70 to 200 billion parameters in parallel, RAG and multi-GPU fine-tuning. Recommended runtimes: vLLM and ninfer (optimized for Blackwell sm_120).

Positioning

RTX PRO 6000 Blackwell server: 96 GB of GDDR7 per card, up to 8 air-cooled cards in a standard chassis, to serve models of 70 to 200 billion parameters.

Manufacturers

Certified servers from Dell, HPE, Lenovo, Supermicro, ASUS and Gigabyte.

GPUDirect Storage

On platforms fitted with ConnectX cards, GPUDirect Storage technology establishes a direct path between NVMe or NVMe over Fabric storage and GPU memory, bypassing the CPU buffer. Compatible arrays, such as NetApp, VAST, DDN or WEKA, feed training and large-scale RAG at full speed.

Support

Open stack, maintained by its communities. Optional NVIDIA AI Enterprise (NIM, NeMo, Triton) with a service-level agreement, licensed per GPU.

Which models fit on RTX PRO 6000 server?

The machine offers 768 GB of memory. 5 of the 7 open models in the catalogue load on it, in the format shown.

Model	Billion parameters	Most precise format that fits	Sizing page
GLM 5.3	744	NVFP4	GLM 5.3 on RTX PRO 6000 server
Kimi K2.7 Code	1,000	NVFP4	Kimi K2.7 Code on RTX PRO 6000 server
Nemotron 3 Ultra	550	FP8	Nemotron 3 Ultra on RTX PRO 6000 server
MiniMax M3	428	FP8	MiniMax M3 on RTX PRO 6000 server
Qwen 3.8 27B	27	FP16	Qwen 3.8 27B on RTX PRO 6000 server

See the full sizing matrix.

Frequently asked questions

How much does a configured rtx pro 6000 cost for an LLM?

Pricing depends on configuration and supplier. No public price is listed for this configuration; a scoping call helps compare supplier quotes.

Which LLM fits in a rtx pro 6000?

Depends on the quantization format and the input context. The table on the page indicates the recommended maximum model size.

Which runtimes support the rtx pro 6000?

vLLM, llama.cpp, Triton Inference Server, depending on the chosen framework. The model, format and GPU combinations measured by QDNA are published in the measurements section.