Skip to main content

NVIDIA RTX 5080 (Blackwell)

Overview

NVIDIA RTX 5080 is the second-tier flagship of the RTX 50 series, launched January 30, 2025, positioned in the mid-to-high range. With 16GB GDDR7 and 10,752 CUDA cores, it delivers approximately 50% of the RTX 5090's performance.

Core Specifications

ItemSpec
ArchitectureBlackwell (GB203)
ProcessTSMC 4N
CUDA Cores10,752
Tensor Cores336 (5th Gen)
RT Cores84 (4th Gen)
Boost Clock2.62 GHz
Memory16 GB GDDR7
Memory Bandwidth960 GB/s (30 Gbps × 256-bit)
FP16 Tensor196 TFLOPS (sparse)
FP8 Tensor392 TFLOPS (sparse)
FP4 Tensor1,568 TOPS (sparse)
TDP360 W
Launch Price$999
Launch Date2025-01-30

Comparison with RTX 4080 SUPER

MetricRTX 5080RTX 4080 SUPERImprovement
CUDA Cores10,75210,2401.05×
Memory16GB GDDR716GB GDDR6XSame capacity / faster
Memory Bandwidth960 GB/s736 GB/s1.30×
FP4 Tensor1,568 TOPSN/ANew
TDP360W320W1.13×
Price$999$999Same

LLM Inference Performance

ModelQuantizationPerformance
Llama 3 8BFP16~50 tok/s
Llama 3 13BINT4~38 tok/s
Mistral 7BFP16~62 tok/s
Mixtral 8x7BINT4~25 tok/s

16GB memory can run 13B quantized models or 7B at full precision.

Vendor Information

ItemDetail
VendorNVIDIA
Launch Price$999
Target MarketMainstream consumer AI, 1080p/1440p gaming, Stable Diffusion