NVIDIA T4

Data center GPU · Turing

Summary

The NVIDIA T4 is a low-power, Turing-generation data centre GPU with 16GB of GDDR6. It was widely used for inference, video transcoding and virtual desktops. Its Ada-generation successor in the same tier is the NVIDIA L4.

Launched
2018
VRAM
16 GB
Mem. bandwidth
320 GB/s
On-demand from
$0.42 /GPU/h
Compare prices from 3 providers →

Tech specs

VRAM 16 GB GDDR6
Memory bandwidth 320 GB/s
Interface PCIe
CUDA cores 2,560
Tensor cores 320 (Gen 2)
TDP 70 W
Supported data types
FP64FP32FP16INT8

Cloud rental prices

3 providers · available in 23 countries · lowest price per GPU, per hour

Prices updated

On-demand
$0.42 – $1.37 /GPU/h
Reserved
$0.18 – $0.86 /GPU/h
Spot
$0.07 – $0.14 /GPU/h
Provider On-demand Reserved Spot
AWS ? $0.42 $0.18 —
Azure ? $0.53 $0.20 $0.07
LeaderGPU ✓ $0.56 $0.25 —
Compare all NVIDIA T4 cloud providers & configurations →

Models that fit in VRAM

Total VRAM From 16-bit inference 8-bit inference 4-bit inference
1× T4 16 GB $0.18/h Qwen3-4B-Instruct-2507Rio-3.0-Open-MiniNVIDIA-Nemotron-3-Nano-4B-BF16 Qwen3-4B-Instruct-2507Rio-3.0-Open-MiniNVIDIA-Nemotron-3-Nano-4B-BF16 gpt-oss-20bQwen3-4B-Instruct-2507Rio-3.0-Open-Mini
2× T4 32 GB $1.19/h Qwen3-4B-Instruct-2507Rio-3.0-Open-MiniNVIDIA-Nemotron-3-Nano-4B-BF16 gpt-oss-20bQwen3-4B-Instruct-2507Rio-3.0-Open-Mini GLM-4.7-FlashQwen3-Coder-30B-A3B-InstructNVIDIA-Nemotron-3-Nano-30B-A3B-BF16
4× T4 64 GB $1.64/h gpt-oss-20bQwen3-4B-Instruct-2507Rio-3.0-Open-Mini GLM-4.7-FlashQwen3-Coder-30B-A3B-InstructNVIDIA-Nemotron-3-Nano-30B-A3B-BF16 GLM-4.5-AirQwen3-Coder-NextNVIDIA-Nemotron-3-Nano-30B-A3B-BF16
6× T4 96 GB $1.51/h GLM-4.7-FlashQwen3-Coder-30B-A3B-InstructNVIDIA-Nemotron-3-Nano-30B-A3B-BF16 Qwen3-Coder-NextGLM-4.7-FlashNVIDIA-Nemotron-3-Nano-30B-A3B-BF16 DeepSeek-V4-Flashgpt-oss-120bNVIDIA-Nemotron-3-Super-120B-A12B-BF16
8× T4 128 GB $3.38/h GLM-4.7-FlashQwen3-Coder-30B-A3B-InstructNVIDIA-Nemotron-3-Nano-30B-A3B-BF16 GLM-4.5-AirQwen3-Coder-NextNVIDIA-Nemotron-3-Nano-30B-A3B-BF16 DeepSeek-V4-Flashgpt-oss-120bNVIDIA-Nemotron-3-Super-120B-A12B-BF16

Open-weights models that fit in VRAM. Estimated using 🤗 accelerate, plus approximation for up to 8K context.

Resources

Detailed documentation from the manufacturer.

Whitepaper ↗ Architecture deep-dive (PDF) Datasheet ↗ Official spec sheet Product page ↗ Manufacturer overview