PNY GeForce RTX 5090 32GB GDDR7

★★★★☆ 3.5 (12)

Equipped with 32GB of GDDR7 memory and 21,760 CUDA cores on the Blackwell architecture, this GPU delivers a 79.6/100 AI performance score for accelerating large language models. Its triple-fan cooler manages a 575W TDP while DLSS 4 and full ray tracing support provide high frame rates at extreme resolutions. This card is best for AI researchers and developers running massive local models who also demand top-tier 4K gaming.

GPU GeForce RTX 5090
VRAM 32GB GDDR7
Boost Clock 2410 MHz
TDP 575 W
Memory Bus 512 bit
Ray Tracing 是
Class consumer
PNY GeForce RTX 5090 32GB GDDR7 gpu
73 综合评分

暂无零售商最新价格

其他可用国家/地区:

概览

优缺点

优点

  • Strong raster (99th percentile) 99th
  • Strong ai (98th percentile) 98th
  • Strong vram (97th percentile) 97th
  • Strong connectivity (94th percentile) 94th

缺点

  • Below average user sentiment (6th percentile)
  • Below average social proof (7th percentile)
  • Below average efficiency (29th percentile)

用户评价

用户口碑如何随时间变化

独家

依据客户实际撰写评价的时间--让你看到最初的好评是否持续。

用户口碑随时间保持稳定
55/100我们的 AI 口碑分析置信度低 · 7 个来源 · 2026年6月
1★2★3★4★5★Q2 '25: 3.4★ · 5 条评价Q4 '25: 3.0★ · 2 条评价52Q2 '25Q4 '25
平均评分满意(4-5★)不满意(1-2★)柱形高度 = 评价数量

基于 7 条带日期的客户评价,按日历季度分组。分期分析为英文。

实测数据

性能表现

性能百分位

AI 97.5
Rt 85.1
Vram 97.4
Raster 99.4
Compute 67.2
Feature 72.7
Thermal 72.9
User Sentiment 6.4
Efficiency 29.1
Connectivity 94
用户口碑 6.5

规格参数

完整规格参数

Chipset

GPU Chipset GeForce RTX 5090
Brand NVIDIA
Architecture Blackwell
Class consumer

Memory

VRAM 32 GB
Memory Type GDDR7
Memory Bus 512
Bandwidth 1792

Compute

CUDA Cores 21760
Tensor Cores 5
RT Cores 4
Base Clock 2010
Boost Clock 2410

Performance

PassMark G3D 38958
GPU Compute 24458

Power & Cooling

TDP 575
Recommended PSU 1000
Power Connectors 1x16 to 4x8 adapter included
Cooler Triple Fan
Fans 3

Connectivity

Interface PCI-Express 5.0 x16
DisplayPort 3
HDMI 1
Max Displays 4
Max Resolution 8K 165Hz

Physical

Length 329
Slot Width 3.5
Weight 2.3 kg / 5.0 lbs

Features

Ray Tracing Yes
Upscaling DLSS 4
AV1 Encode Yes

竞品对比

Spec PNY GeForce RTX 5090 ASUS ProArt RTX 5080 Gigabyte AERO GV-N5080AERO OC-16GD MSI Ventus RTX 5070 Ti 16G VENTUS 3X PZ OC ZOTAC AMP Extreme Infinity ZT-B50800B-10P Sapphire NITRO+ 11322-01-40G
Chipset GeForce RTX 5090 GeForce RTX 5080 GeForce RTX 5080 GeForce RTX 5070 Ti GeForce RTX 5080 Radeon RX 7900 XTX
Vram 32GB GDDR7 16GB GDDR7 16GB GDDR7 16GB GDDR7 16GB GDDR7 24GB GDDR6
Boost Clock Mhz 2410 2730 2730 2482 2670 2680
Tdp W 575 150 360 300 360 339
Memory Bus Bits 512 256 256 256 256 384
Ray Tracing true true true true true true
GPU Class consumer consumer consumer consumer consumer consumer
Compare Compare Compare Compare Compare
产品 AIRtVramRasterComputeFeatureThermalUser SentimentEfficiencyConnectivity用户口碑
PNY GeForce RTX 5090 97.585.197.499.467.272.772.96.429.1946.5
ASUS ProArt RTX 5080 Compare 86.795.888.394.697.296.491.5097.498.695.3
Gigabyte AERO GV-N5080AERO OC-16GD Compare 86.795.888.394.697.296.491.571.857.69488.6
MSI Ventus RTX 5070 Ti 16G VENTUS 3X PZ OC Compare 80.692.585.188.89396.491.580.468.99488.6
ZOTAC AMP Extreme Infinity ZT-B50800B-10P Compare 86.797.688.394.697.296.491.571.857.69499
Sapphire NITRO+ 11322-01-40G Compare 94.691.791.483.499.284.391.586.542.598.778

使用场景评分

总体 (73.1)AI/LLM (87.4)Budget (69.9)Gaming 4k (85.8)Ray Tracing (82.9)Workstation (70.6)Gaming 1080p (74.4)Gaming 1440p (79.2)Content Creation (82.4)

完整评测即将发布。上方的规格、评分、价格和图表始终保持最新。

类似产品