RTX PRO 6000 Blackwell Workstation Edition

Name: RTX PRO 6000 Blackwell Workstation Edition
Brand: NVIDIA
Price: 17999 USD

NVIDIA RTX PRO 6000 Blackwell Workstation Edition: 96 GB of memory at 1792 GB/s. Below: the open models it can run, estimated at 96 GB with an 8K context. 96 GB of VRAM at 1,792 GB/s on one card: 70B models fit at 8-bit, and 120B-class MoE models at 5- to 6-bit.

Last updated: 2026-10-14

Specifications

Memory	96 GB GDDR7 (ECC)[src]
Memory bandwidth	1792 GB/s[src]
Power	600 W[src]
Launch	2025-04[src]
Launch price	US$8,565[src]
Architecture	Blackwell (GB202)[src]
Software	CUDA

What it can run

Best-quality quantization that fits with a 8K context, and the estimated generation speed. Estimates, not benchmarks. See the methodology.

Model	Params	Best quant	Memory	tok/s (est.)
Qwen3.5-122B-A10B	122B (10B)	Q5_K_M	85.5 GiB	176
Mistral Small 4 119B (2603)	119B (6.5B)	Q5_K_M	83.4 GiB	271
gpt-oss-120b	117B (5.1B)	Q6_K	95.0 GiB	299
Llama 4 Scout 17B-16E Instruct	109B (17B)	Q6_K	89.8 GiB	90
GLM-4.5-Air	106B (12B)	Q6_K	87.3 GiB	127
Qwen3-Coder-Next (80B-A3B)	80B (3B)	Q8_0	83.8 GiB	394
Llama 3.3 70B Instruct	70B	Q8_0	75.7 GiB	17
Qwen3.5-35B-A3B	35B (3B)	FP16	69.1 GiB	209
Qwen3.6-35B-A3B	35B (3B)	FP16	69.1 GiB	209
Qwen3-32B	32.8B	FP16	66.6 GiB	19
DeepSeek-R1-Distill-Qwen-32B	32.76B	FP16	66.6 GiB	19
Gemma 4 31B IT	30.7B	FP16	61.6 GiB	20
Qwen3-30B-A3B (2507)	30.5B (3.3B)	FP16	60.9 GiB	190
Qwen3-Coder-30B-A3B-Instruct	30.5B (3.3B)	FP16	60.9 GiB	190
GLM-4.7-Flash (30B-A3B)	30B (3B)	FP16	59.6 GiB	209
Qwen3.5-27B	27B	FP16	53.8 GiB	23
Qwen3.6-27B	27B	FP16	53.8 GiB	23
Qwen3.8-27B	27B	FP16	53.8 GiB	23
Gemma 3 27B IT	27B	FP16	54.3 GiB	23
Gemma 4 26B A4B IT	25.2B (3.8B)	FP16	50.1 GiB	165
Mistral Small 3.2 24B Instruct (2506)	24B	FP16	48.7 GiB	26
Magistral Small 1.2 (2509)	24B	FP16	48.7 GiB	26
Devstral Small 2 24B Instruct (2512)	24B	FP16	48.7 GiB	26
gpt-oss-20b	21B (3.6B)	FP16	41.8 GiB	174
Phi-4 (14B)	14B	FP16	29.4 GiB	45
Gemma 3 12B IT	12B	FP16	24.8 GiB	52
Gemma 4 12B Unified IT	11.95B	FP16	24.2 GiB	52
Qwen3.5-9B	9B	FP16	18.4 GiB	70
Qwen3-8B	8.2B	FP16	17.7 GiB	76
DeepSeek-R1-0528-Qwen3-8B	8.19B	FP16	17.6 GiB	77
Llama 3.1 8B Instruct	8B	FP16	17.1 GiB	78
Gemma 4 E4B IT	8B	FP16	16.4 GiB	78
Llama 3.2 3B Instruct	3.21B	FP16	7.7 GiB	195
Kimi K2.6	1000B (32B)	Doesn't fit	–	–
DeepSeek-R1-0528 (671B)	671B (37B)	Doesn't fit	–	–
DeepSeek-V4-Flash	284B (13B)	Doesn't fit	–	–
Qwen3.8-Flash-Next	180B (6B)	Doesn't fit	–	–

Prices

Latest price at each shop, as listed on the day we checked.

Region	Shop	Listing	Price	Per GB	Checked
US	Newegg	NVIDIA RTX PRO 6000 Blackwell Workstation Edition, 96 GB GDDR7 with ECC, PCI Express 5.0 x16, OEM packageOEM package; $19.99 shipping not included	US$17,999	US$187/GB	2026-10-14

US value: about 258.6 tok/s on Llama 3.1 8B Instruct (Q4_K_M, 8K), or 1.4 tok/s per US$100.

A price history chart appears once we have checked this device on two or more days.

All prices →

Frequently asked questions

Which LLMs can the RTX PRO 6000 Blackwell Workstation Edition run?

With 96 GB and an 8K context, the largest models that fit include: Qwen3.5-122B-A10B, Mistral Small 4 119B (2603), gpt-oss-120b, Llama 4 Scout 17B-16E Instruct, GLM-4.5-Air. See the full table below.

How fast is the RTX PRO 6000 Blackwell Workstation Edition for local LLMs?

Generation speed is mostly set by memory bandwidth (1792 GB/s). For Llama 3.1 8B Instruct at FP16, expect roughly 78 tokens/s; bigger models are slower.

Can the RTX PRO 6000 Blackwell Workstation Edition run 70B models?

Yes. Llama 3.3 70B Instruct fits on the RTX PRO 6000 Blackwell Workstation Edition (96 GB) at Q8_0 with an 8K context, using about 75.7 GiB, at an estimated 17 tokens/s.