Skip to content
Updated minutes ago
nvidia

RTX 4080

nvidia · ada · 16 GB GDDR6X · 320W TDP

VRAM

16 GB

BF16 TFLOPS

97

Bandwidth

717 GB/s

From

$0.32/hr

Calculate ROI with this GPU →

Spec Sheet

VRAM16 GB GDDR6X
Memory Bandwidth717 GB/s
BF16 TFLOPS97
FP16 TFLOPS97
FP8 TFLOPS194
INT8 TOPS194
TDP320W
InterconnectPCIE
Max per Node4
PCIe Gen4
CUDA Compute Capability8.9
Tensor CoresYes

Pricing by Provider

ProviderOn-DemandReservedSpotBadge
tensordock$0.49/hr-$0.32/hrCheapest
vast_ai$0.55/hr-$0.35/hr

Compatible Models (233)

Training Capabilities

Estimated GPU count for full fine-tuning (AdamW, BF16) and QLoRA

Model SizeFull Fine-TuneQLoRA
7B model9 GPUs1 GPU
13B model16 GPUs1 GPU
70B model83 GPUs3 GPUs

Energy Efficiency

Estimated tokens/second per Watt for popular models

Mistral 7B
0.31 t/s/WFP8
Qwen 2.5 7B
0.29 t/s/WFP8
Llama 3.1 8B
0.28 t/s/WFP8

Similar GPUs

GPUVRAMBF16 TFLOPSBW (GB/s)From
RTX 4060 Ti 16GB16 GB44288$0.30/hr
RTX 4070 Ti12 GB93504$0.25/hr
RTX 4070 Super12 GB55504$0.22/hr
L424 GB121300$0.29/hr
RTX 409024 GB1651008$0.39/hr

Methodology Note

Performance estimates for the RTX 4080are based on InferenceBench's roofline performance model with CUDA kernel-level optimization including FlashAttention v2 and PagedAttention. Memory calculations account for model weights (16 GB GDDR6X available), KV-cache allocation, and activation memory. Throughput predictions use the RTX 4080's rated 717 GB/s memory bandwidth and 97 BF16 TFLOPS compute capacity as roofline ceilings, with empirical correction factors per GPU architecture (ada). See our full methodology.

Frequently Asked Questions

How many AI models can run on RTX 4080?

The RTX 4080 can run 233 AI models from our database within a single node. Compatible models range across various parameter sizes depending on the quantization precision (BF16, FP8, INT4). Smaller models fit on a single GPU while larger models may require multi-GPU setups up to 4x RTX 4080.

What is the RTX 4080 inference throughput?

The RTX 4080 delivers 97 BF16 TFLOPS and 194 FP8 TFLOPS with 717 GB/s memory bandwidth. Actual inference throughput (tokens/sec) depends on the model size, precision, and batch size. Use our calculator for model-specific throughput estimates.

How much does RTX 4080 cost per hour?

The RTX 4080 is available starting from $0.32/hour via tensordock. Prices vary by provider and pricing tier (on-demand, reserved, spot). Compare pricing across all providers in the table above.