Skip to content

NVIDIA · Blackwell 2.0 · 96 GB

RTX PRO 6000 Blackwell

96 GB GDDR7 · 512-bit · 24064 shaders · 600 W

1080p
103%
1440p
106%
4K
110%
Amazon
$15500
$15500
Available Updated 3 days, 3 hours ago
Gaming 1080p103.1%
Gaming 1440p106.1%
Gaming 4K110.1%
1440p value6.8

Written analysis

RTX PRO 6000 Blackwell at a glance

The RTX PRO 6000 Blackwell is based on the Blackwell 2.0 architecture and is manufactured using the 5 nm process at TSMC. It ships with 96 GB of GDDR7 on a 512-bit bus , about 1 GB/s. Shader count is 24064 across 188 CU/SM. The board interface is PCIe 5.0 x16.

Reference board power is a very high 600 W with 1x 16-pin. PSU sizing must include the CPU and transient load, while the case needs unrestricted intake and exhaust.

VRAM capacity is high-end at 96 GB. That is the class you want for 4K textures, heavy ray tracing, and creator GPU memory loads.

Best fit: High-refresh 1440p and 4K-oriented systems that can support its power and cooling requirements.

Gaming performance

RTX PRO 6000 Blackwell is a high-end gaming GPU, at 106.10% of the dataset’s best 1440p raster performance. Pair it with a fast CPU and a high-refresh 1440p or 4K display if you want that headroom.

1440p value is 6.8 performance per dollar. You are paying for peak raster performance, not efficiency.

The catalog lists 1 retail card of this GPU . Partner cards change cooling, length, factory clocks, and price — not the core raster scores. They are listed below similar GPUs.

Silicon

Full specifications

Architecture, memory subsystem, clocks, compute resources, throughput, power, and API support in one technical sheet.

Identity

GPU
RTX PRO 6000 Blackwell
Chip name
GB202
Launch price
$8565
Amazon price
$15500
Availability
In stock
Price updated
3 days, 3 hours ago
Launch
1 year, 5 months ago

Architecture

Architecture
Blackwell 2.0
Foundry
TSMC
Process
5 nm
Transistors
92200 million
Die size
750 mm²
Interface
PCIe 5.0 x16

Memory

Capacity / type
96 GB GDDR7
Bus width
512-bit
Bandwidth
1 GB/s
Memory clock
1750 MHz
L2 cache
128 MB

Clocks

Base clock
1590 MHz
Boost clock
2617 MHz

Compute

Shader cores
24064
CU / SM
188
Texture units
752
ROPs
192
RT cores
188
Tensor / matrix
752

Throughput

Pixel rate
502.5 GPixel/s
Texture rate
1968.0 GTexel/s
FP32
126.00 TFLOPS
FP16
126.01 TFLOPS
FP64
1.97 TFLOPS

Power

Board power
600 W
Connectors
1x 16-pin
Reference slots
Dual-slot

APIs

DirectX
12 Ultimate (12_2)
Vulkan
1.4
OpenGL
4.6
OpenCL
3.0
Shader model
6.8

Amazon

Price history

Nearby performance

Similar GPUs

Silicon close to RTX PRO 6000 Blackwell on the same 1440p scale. Open the GPU table to sort, filter, and compare side by side.

NVIDIA RTX 5090 32 GB GDDR7 · Blackwell 2.0
100.0%
100.0%
19.0
Compare these in the GPU table

Partner boards

RTX PRO 6000 Blackwell graphics cards

Every retail card of this GPU in the catalog. Length, cooler, and price change; raster scores stay the same.

Open these in the graphics card table

Discussion

Comments on RTX PRO 6000 Blackwell

@Ivica 1 year ago

The RTX 6000 (Blackwell) is currently the fastest GPU available for 4K performance and compute-heavy tasks, outperforming even the RTX 5090 by around 10% in ultra-high-resolution scenarios. However, it’s important to note that this is a workstation-class GPU, not designed for mainstream gaming use. With a price tag close to $9,000 USD, it's targeted at professionals working in rendering, simulation, AI, and data science—not typical consumers or gamers.

For users who want the absolute best performance regardless of cost, this GPU is unmatched. But for nearly all other users, especially gamers, the RTX 5090 remains the top-performing consumer graphics card and the benchmark we use for 100% relative GPU performance on this s