Skip to content

Hugging Face

28 offerings · 56 billing options

Products: Inference Endpoints

0 of 28 offerings verified recently

7 identified GPU models

View official pricing source →

Dedicated GPU prices

1× A100 · 80 GB VRAM/GPU
AWS NVIDIA A100 x1 · Dedicated GPU · fixed
  • USD 2.5/hour
    compute time
≈ EUR 2.177/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
1× A100 · 80 GB VRAM/GPU
AWS NVIDIA A100 x1 · Dedicated GPU · fixed
  • USD 2.5/hour
    compute time
≈ EUR 2.177/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
2× A100 · 80 GB VRAM/GPU
AWS NVIDIA A100 x2 · Dedicated GPU · fixed
  • USD 5/hour
    compute time
≈ EUR 4.354/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
2× A100 · 80 GB VRAM/GPU
AWS NVIDIA A100 x2 · Dedicated GPU · fixed
  • USD 5/hour
    compute time
≈ EUR 4.354/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
4× A100 · 80 GB VRAM/GPU
AWS NVIDIA A100 x4 · Dedicated GPU · fixed
  • USD 10/hour
    compute time
≈ EUR 8.707/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
4× A100 · 80 GB VRAM/GPU
AWS NVIDIA A100 x4 · Dedicated GPU · fixed
  • USD 10/hour
    compute time
≈ EUR 8.707/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
8× A100 · 80 GB VRAM/GPU
AWS NVIDIA A100 x8 · Dedicated GPU · fixed
  • USD 20/hour
    compute time
≈ EUR 17.41/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
8× A100 · 80 GB VRAM/GPU
AWS NVIDIA A100 x8 · Dedicated GPU · fixed
  • USD 20/hour
    compute time
≈ EUR 17.41/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
1× A10G · 24 GB VRAM/GPU
AWS NVIDIA A10G x1 · Dedicated GPU · fixed
  • USD 1/hour
    compute time
≈ EUR 0.8707/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
1× A10G · 24 GB VRAM/GPU
AWS NVIDIA A10G x1 · Dedicated GPU · fixed
  • USD 1/hour
    compute time
≈ EUR 0.8707/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
4× A10G · 24 GB VRAM/GPU
AWS NVIDIA A10G x4 · Dedicated GPU · fixed
  • USD 5/hour
    compute time
≈ EUR 4.354/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
4× A10G · 24 GB VRAM/GPU
AWS NVIDIA A10G x4 · Dedicated GPU · fixed
  • USD 5/hour
    compute time
≈ EUR 4.354/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
1× NVIDIA H200 · 141 GB VRAM/GPU
AWS NVIDIA H200 x1 · Dedicated GPU · fixed
  • USD 5/hour
    compute time
≈ EUR 4.354/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
1× NVIDIA H200 · 141 GB VRAM/GPU
AWS NVIDIA H200 x1 · Dedicated GPU · fixed
  • USD 5/hour
    compute time
≈ EUR 4.354/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
2× NVIDIA H200 · 141 GB VRAM/GPU
AWS NVIDIA H200 x2 · Dedicated GPU · fixed
  • USD 10/hour
    compute time
≈ EUR 8.707/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
2× NVIDIA H200 · 141 GB VRAM/GPU
AWS NVIDIA H200 x2 · Dedicated GPU · fixed
  • USD 10/hour
    compute time
≈ EUR 8.707/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
4× NVIDIA H200 · 141 GB VRAM/GPU
AWS NVIDIA H200 x4 · Dedicated GPU · fixed
  • USD 20/hour
    compute time
≈ EUR 17.41/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
4× NVIDIA H200 · 141 GB VRAM/GPU
AWS NVIDIA H200 x4 · Dedicated GPU · fixed
  • USD 20/hour
    compute time
≈ EUR 17.41/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
8× NVIDIA H200 · 141 GB VRAM/GPU
AWS NVIDIA H200 x8 · Dedicated GPU · fixed
  • USD 40/hour
    compute time
≈ EUR 34.83/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
8× NVIDIA H200 · 141 GB VRAM/GPU
AWS NVIDIA H200 x8 · Dedicated GPU · fixed
  • USD 40/hour
    compute time
≈ EUR 34.83/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
1× L4 · 24 GB VRAM/GPU
AWS NVIDIA L4 x1 · Dedicated GPU · fixed
  • USD 0.8/hour
    compute time
≈ EUR 0.6966/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
1× L4 · 24 GB VRAM/GPU
AWS NVIDIA L4 x1 · Dedicated GPU · fixed
  • USD 0.8/hour
    compute time
≈ EUR 0.6966/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
4× L4 · 24 GB VRAM/GPU
AWS NVIDIA L4 x4 · Dedicated GPU · fixed
  • USD 3.8/hour
    compute time
≈ EUR 3.309/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
4× L4 · 24 GB VRAM/GPU
AWS NVIDIA L4 x4 · Dedicated GPU · fixed
  • USD 3.8/hour
    compute time
≈ EUR 3.309/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
1× L40S · 48 GB VRAM/GPU
AWS NVIDIA L40S x1 · Dedicated GPU · fixed
  • USD 1.8/hour
    compute time
≈ EUR 1.567/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
1× L40S · 48 GB VRAM/GPU
AWS NVIDIA L40S x1 · Dedicated GPU · fixed
  • USD 1.8/hour
    compute time
≈ EUR 1.567/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
4× L40S · 48 GB VRAM/GPU
AWS NVIDIA L40S x4 · Dedicated GPU · fixed
  • USD 8.3/hour
    compute time
≈ EUR 7.227/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
4× L40S · 48 GB VRAM/GPU
AWS NVIDIA L40S x4 · Dedicated GPU · fixed
  • USD 8.3/hour
    compute time
≈ EUR 7.227/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
8× L40S · 48 GB VRAM/GPU
AWS NVIDIA L40S x8 · Dedicated GPU · fixed
  • USD 23.5/hour
    compute time
≈ EUR 20.46/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
8× L40S · 48 GB VRAM/GPU
AWS NVIDIA L40S x8 · Dedicated GPU · fixed
  • USD 23.5/hour
    compute time
≈ EUR 20.46/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
1× T4 · 14 GB VRAM/GPU
AWS NVIDIA T4 x1 · Dedicated GPU · fixed
  • USD 0.5/hour
    compute time
≈ EUR 0.4354/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
1× T4 · 14 GB VRAM/GPU
AWS NVIDIA T4 x1 · Dedicated GPU · fixed
  • USD 0.5/hour
    compute time
≈ EUR 0.4354/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
4× T4 · 14 GB VRAM/GPU
AWS NVIDIA T4 x4 · Dedicated GPU · fixed
  • USD 3/hour
    compute time
≈ EUR 2.612/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
4× T4 · 14 GB VRAM/GPU
AWS NVIDIA T4 x4 · Dedicated GPU · fixed
  • USD 3/hour
    compute time
≈ EUR 2.612/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
1× A100 · 80 GB VRAM/GPU
GCP NVIDIA A100 x1 · Dedicated GPU · fixed
  • USD 3.6/hour
    compute time
≈ EUR 3.135/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
1× A100 · 80 GB VRAM/GPU
GCP NVIDIA A100 x1 · Dedicated GPU · fixed
  • USD 3.6/hour
    compute time
≈ EUR 3.135/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
2× A100 · 80 GB VRAM/GPU
GCP NVIDIA A100 x2 · Dedicated GPU · fixed
  • USD 7.2/hour
    compute time
≈ EUR 6.269/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
2× A100 · 80 GB VRAM/GPU
GCP NVIDIA A100 x2 · Dedicated GPU · fixed
  • USD 7.2/hour
    compute time
≈ EUR 6.269/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
4× A100 · 80 GB VRAM/GPU
GCP NVIDIA A100 x4 · Dedicated GPU · fixed
  • USD 14.4/hour
    compute time
≈ EUR 12.54/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
4× A100 · 80 GB VRAM/GPU
GCP NVIDIA A100 x4 · Dedicated GPU · fixed
  • USD 14.4/hour
    compute time
≈ EUR 12.54/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
8× A100 · 80 GB VRAM/GPU
GCP NVIDIA A100 x8 · Dedicated GPU · fixed
  • USD 28.8/hour
    compute time
≈ EUR 25.08/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
8× A100 · 80 GB VRAM/GPU
GCP NVIDIA A100 x8 · Dedicated GPU · fixed
  • USD 28.8/hour
    compute time
≈ EUR 25.08/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
1× H100 · 80 GB VRAM/GPU
GCP NVIDIA H100 x1 · Dedicated GPU · fixed
  • USD 10/hour
    compute time
≈ EUR 8.707/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
1× H100 · 80 GB VRAM/GPU
GCP NVIDIA H100 x1 · Dedicated GPU · fixed
  • USD 10/hour
    compute time
≈ EUR 8.707/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
2× H100 · 80 GB VRAM/GPU
GCP NVIDIA H100 x2 · Dedicated GPU · fixed
  • USD 20/hour
    compute time
≈ EUR 17.41/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
2× H100 · 80 GB VRAM/GPU
GCP NVIDIA H100 x2 · Dedicated GPU · fixed
  • USD 20/hour
    compute time
≈ EUR 17.41/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
4× H100 · 80 GB VRAM/GPU
GCP NVIDIA H100 x4 · Dedicated GPU · fixed
  • USD 40/hour
    compute time
≈ EUR 34.83/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
4× H100 · 80 GB VRAM/GPU
GCP NVIDIA H100 x4 · Dedicated GPU · fixed
  • USD 40/hour
    compute time
≈ EUR 34.83/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
8× H100 · 80 GB VRAM/GPU
GCP NVIDIA H100 x8 · Dedicated GPU · fixed
  • USD 80/hour
    compute time
≈ EUR 69.66/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
8× H100 · 80 GB VRAM/GPU
GCP NVIDIA H100 x8 · Dedicated GPU · fixed
  • USD 80/hour
    compute time
≈ EUR 69.66/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
1× L4 · 24 GB VRAM/GPU
GCP NVIDIA L4 x1 · Dedicated GPU · fixed
  • USD 0.7/hour
    compute time
≈ EUR 0.6095/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
1× L4 · 24 GB VRAM/GPU
GCP NVIDIA L4 x1 · Dedicated GPU · fixed
  • USD 0.7/hour
    compute time
≈ EUR 0.6095/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
4× L4 · 24 GB VRAM/GPU
GCP NVIDIA L4 x4 · Dedicated GPU · fixed
  • USD 3.8/hour
    compute time
≈ EUR 3.309/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
4× L4 · 24 GB VRAM/GPU
GCP NVIDIA L4 x4 · Dedicated GPU · fixed
  • USD 3.8/hour
    compute time
≈ EUR 3.309/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
1× T4 · 16 GB VRAM/GPU
GCP NVIDIA T4 x1 · Dedicated GPU · fixed
  • USD 0.5/hour
    compute time
≈ EUR 0.4354/hour
Always on
always on · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.
1× T4 · 16 GB VRAM/GPU
GCP NVIDIA T4 x1 · Dedicated GPU · fixed
  • USD 0.5/hour
    compute time
≈ EUR 0.4354/hour
Scale to zero
scale to zero · dedicated
1 minute increments · 1 minute minimum · bills initializing, running
GPU rows do not publish system memory, so systemMemoryGb remains null.

Other GPU Providers