---
title: Cloud GPU Prices catalog results
description: 83 matching offerings and 124 commercial variants
manual: https://cloudgpuprices.com/agents
api: https://cloudgpuprices.com/api/v1/offerings?category=serverless_gpu&min_gpu_memory_gb=24
---

# Cloud GPU Prices catalog results

Found **83 offerings** and **124 commercial variants**. This page contains 50 offerings starting at offset 0.

Read the [agent and data access manual](https://cloudgpuprices.com/agents) before interpreting or comparing prices. In particular, `unknown` does not mean `no`, and only rows marked as complete total prices may be ranked as totals.

Exact structured data: [JSON API](https://cloudgpuprices.com/api/v1/offerings?category=serverless_gpu&min_gpu_memory_gb=24)

| Provider | Offering | GPU | VRAM/GPU | Region | Variant | Effective rate | Complete total | EU compute | EU-constrained | Freshness |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Baseten.co | A10G | A10G | 24 GB | unknown | Standard | USD 0.02012/minute | no | unknown | unknown | stale |
| Baseten.co | L4 | L4 | 24 GB | unknown | Standard | USD 0.01414/minute | no | unknown | unknown | stale |
| Baseten.co | B200 | NVIDIA B200 | 180 GB | unknown | Standard | USD 0.16633/minute | no | unknown | unknown | stale |
| Baseten.co | H100 MIG | not guaranteed | 40 GB | unknown | Standard | USD 0.0625/minute | no | unknown | unknown | stale |
| Baseten.co | H100 | H100 | 80 GB | unknown | Standard | USD 0.10833/minute | no | unknown | unknown | stale |
| Baseten.co | A100 | A100 | 80 GB | unknown | Standard | USD 0.06667/minute | no | unknown | unknown | stale |
| Beam.cloud | B200 | NVIDIA B200 | 180 GB | unknown | Standard | USD 0.000013/core_second + USD 0.001561/second + USD 0.000002/gib_second | no | unknown | unknown | stale |
| Beam.cloud | RTX 5090 | NVIDIA GeForce RTX 5090 | 32 GB | unknown | Standard | USD 0.000013/core_second + USD 0.000303/second + USD 0.000002/gib_second | no | unknown | unknown | stale |
| Beam.cloud | L40S | L40S | 48 GB | unknown | Standard | USD 0.000013/core_second + USD 0.000486/second + USD 0.000002/gib_second | no | unknown | unknown | stale |
| Beam.cloud | A100 80GB | A100 | 80 GB | unknown | Standard | USD 0.000013/core_second + USD 0.000625/second + USD 0.000002/gib_second | no | unknown | unknown | stale |
| Beam.cloud | H100 | H100 | 80 GB | unknown | Standard | USD 0.000013/core_second + USD 0.000986/second + USD 0.000002/gib_second | no | unknown | unknown | stale |
| Beam.cloud | RTX PRO 6000 | NVIDIA RTX PRO 6000 Blackwell | 96 GB | unknown | Standard | USD 0.000013/core_second + USD 0.000758/second + USD 0.000002/gib_second | no | unknown | unknown | stale |
| Beam.cloud | RTX 4090 | RTX 4090 | 24 GB | unknown | Standard | USD 0.000013/core_second + USD 0.000192/second + USD 0.000002/gib_second | no | unknown | unknown | stale |
| Beam.cloud | H200 | NVIDIA H200 | 141 GB | unknown | Standard | USD 0.000013/core_second + USD 0.001136/second + USD 0.000002/gib_second | no | unknown | unknown | stale |
| Beam.cloud | A6000 | A6000 | 48 GB | unknown | Standard | USD 0.000013/core_second + USD 0.000228/second + USD 0.000002/gib_second | no | unknown | unknown | stale |
| Cerebrium | A100 (40GB) | A100 | 40 GB | unknown | Standard | USD 0.000007/core_second + USD 0.000555/second + USD 0.000002/gb_second | no | yes | yes | stale |
| Cerebrium | A100 (80GB) | A100 | 80 GB | unknown | Standard | USD 0.000007/core_second + USD 0.000583/second + USD 0.000002/gb_second | no | yes | yes | stale |
| Fal.ai | B200 | NVIDIA B200 | 180 GB | unknown | B200 always-on | USD 6.25/hour | yes | unknown | unknown | stale |
| Fal.ai | B200 | NVIDIA B200 | 180 GB | unknown | B200 scale-to-zero | USD 6.25/hour | yes | unknown | unknown | stale |
| Fal.ai | B300 | NVIDIA B300 | 288 GB | unknown | B300 always-on | USD 8.5/hour | yes | unknown | unknown | stale |
| Fal.ai | B300 | NVIDIA B300 | 288 GB | unknown | B300 scale-to-zero | USD 8.5/hour | yes | unknown | unknown | stale |
| Fal.ai | H100 | H100 | 80 GB | unknown | H100 always-on | USD 3.99/hour | yes | unknown | unknown | stale |
| Fal.ai | H100 | H100 | 80 GB | unknown | H100 scale-to-zero | USD 3.99/hour | yes | unknown | unknown | stale |
| Fal.ai | RTX PRO 6000 | NVIDIA RTX PRO 6000 Blackwell | 96 GB | unknown | RTX PRO 6000 always-on | USD 2.99/hour | yes | unknown | unknown | stale |
| Fal.ai | RTX PRO 6000 | NVIDIA RTX PRO 6000 Blackwell | 96 GB | unknown | RTX PRO 6000 scale-to-zero | USD 2.99/hour | yes | unknown | unknown | stale |
| Fal.ai | H200 | NVIDIA H200 | 141 GB | unknown | H200 always-on | USD 4.5/hour | yes | unknown | unknown | stale |
| Fal.ai | H200 | NVIDIA H200 | 141 GB | unknown | H200 scale-to-zero | USD 4.5/hour | yes | unknown | unknown | stale |
| Google Cloud | NVIDIA-L4 | L4 | 24 GB | unknown | NVIDIA-L4 No zonal redundancy | USD 0.000018/core_second + USD 0.000187/second + USD 0.000002/gib_second | no | yes | yes | stale |
| Google Cloud | NVIDIA-L4 | L4 | 24 GB | unknown | NVIDIA-L4 Zonal redundancy | USD 0.000018/core_second + USD 0.000291/second + USD 0.000002/gib_second | no | yes | yes | stale |
| Google Cloud | NVIDIA RTX Pro 6000 | NVIDIA RTX PRO 6000 Blackwell | 96 GB | unknown | NVIDIA RTX Pro 6000 No zonal redundancy | USD 0.000018/core_second + USD 0.000365/second + USD 0.000002/gib_second | no | yes | yes | stale |
| Google Cloud | NVIDIA RTX Pro 6000 | NVIDIA RTX PRO 6000 Blackwell | 96 GB | unknown | NVIDIA RTX Pro 6000 Zonal redundancy | USD 0.000018/core_second + USD 0.000569/second + USD 0.000002/gib_second | no | yes | yes | stale |
| Inferless | Nvidia A10 Dedicated | A10 | 24 GB | unknown | Standard | USD 0.000341/second | yes | unknown | unknown | stale |
| Inferless | Nvidia A100 Dedicated | A100 | 80 GB | unknown | Standard | USD 0.001491/second | yes | unknown | unknown | stale |
| Inferless | Nvidia A100 Shared | A100 | 40 GB | unknown | Standard | USD 0.000745/second | yes | unknown | unknown | stale |
| Koyeb | L40S | L40S | 48 GB | unknown | Always-on | USD 1.2/hour | yes | unknown | unknown | stale |
| Koyeb | L40S | L40S | 48 GB | unknown | Scale-to-Zero | USD 1.2/hour | yes | unknown | unknown | stale |
| Koyeb | A100 | A100 | 80 GB | unknown | Always-on | USD 1.6/hour | yes | unknown | unknown | stale |
| Koyeb | A100 | A100 | 80 GB | unknown | Scale-to-Zero | USD 1.6/hour | yes | unknown | unknown | stale |
| Koyeb | 2x H200 | NVIDIA H200 | 141 GB | unknown | Always-on | USD 6/hour | yes | unknown | unknown | stale |
| Koyeb | 2x H200 | NVIDIA H200 | 141 GB | unknown | Scale-to-Zero | USD 6/hour | yes | unknown | unknown | stale |
| Koyeb | 8x A100 | A100 | 80 GB | unknown | Always-on | USD 12.8/hour | yes | unknown | unknown | stale |
| Koyeb | 8x A100 | A100 | 80 GB | unknown | Scale-to-Zero | USD 12.8/hour | yes | unknown | unknown | stale |
| Koyeb | 4x H100 | H100 | 80 GB | unknown | Always-on | USD 10/hour | yes | unknown | unknown | stale |
| Koyeb | 4x H100 | H100 | 80 GB | unknown | Scale-to-Zero | USD 10/hour | yes | unknown | unknown | stale |
| Koyeb | L4 | L4 | 24 GB | unknown | Always-on | USD 0.7/hour | yes | unknown | unknown | stale |
| Koyeb | L4 | L4 | 24 GB | unknown | Scale-to-Zero | USD 0.7/hour | yes | unknown | unknown | stale |
| Koyeb | 8x H100 | H100 | 80 GB | unknown | Always-on | USD 20/hour | yes | unknown | unknown | stale |
| Koyeb | 8x H100 | H100 | 80 GB | unknown | Scale-to-Zero | USD 20/hour | yes | unknown | unknown | stale |
| Koyeb | A100 SXM | A100 | 80 GB | unknown | Always-on | USD 2.15/hour | yes | unknown | unknown | stale |
| Koyeb | A100 SXM | A100 | 80 GB | unknown | Scale-to-Zero | USD 2.15/hour | yes | unknown | unknown | stale |
| Koyeb | RTX-PRO-6000 | NVIDIA RTX PRO 6000 Blackwell | 96 GB | unknown | Always-on | USD 2.2/hour | yes | unknown | unknown | stale |
| Koyeb | RTX-PRO-6000 | NVIDIA RTX PRO 6000 Blackwell | 96 GB | unknown | Scale-to-Zero | USD 2.2/hour | yes | unknown | unknown | stale |
| Koyeb | 4x A100 SXM | A100 | 80 GB | unknown | Always-on | USD 8.6/hour | yes | unknown | unknown | stale |
| Koyeb | 4x A100 SXM | A100 | 80 GB | unknown | Scale-to-Zero | USD 8.6/hour | yes | unknown | unknown | stale |
| Koyeb | 4x A100 | A100 | 80 GB | unknown | Always-on | USD 6.4/hour | yes | unknown | unknown | stale |
| Koyeb | 4x A100 | A100 | 80 GB | unknown | Scale-to-Zero | USD 6.4/hour | yes | unknown | unknown | stale |
| Koyeb | 8x H200 | NVIDIA H200 | 141 GB | unknown | Always-on | USD 24/hour | yes | unknown | unknown | stale |
| Koyeb | 8x H200 | NVIDIA H200 | 141 GB | unknown | Scale-to-Zero | USD 24/hour | yes | unknown | unknown | stale |
| Koyeb | H100 | H100 | 80 GB | unknown | Always-on | USD 2.5/hour | yes | unknown | unknown | stale |
| Koyeb | H100 | H100 | 80 GB | unknown | Scale-to-Zero | USD 2.5/hour | yes | unknown | unknown | stale |
| Koyeb | 2x H100 | H100 | 80 GB | unknown | Always-on | USD 5/hour | yes | unknown | unknown | stale |
| Koyeb | 2x H100 | H100 | 80 GB | unknown | Scale-to-Zero | USD 5/hour | yes | unknown | unknown | stale |
| Koyeb | 8x A100 SXM | A100 | 80 GB | unknown | Always-on | USD 17.2/hour | yes | unknown | unknown | stale |
| Koyeb | 8x A100 SXM | A100 | 80 GB | unknown | Scale-to-Zero | USD 17.2/hour | yes | unknown | unknown | stale |
| Koyeb | 2x A100 | A100 | 80 GB | unknown | Always-on | USD 3.2/hour | yes | unknown | unknown | stale |
| Koyeb | 2x A100 | A100 | 80 GB | unknown | Scale-to-Zero | USD 3.2/hour | yes | unknown | unknown | stale |
| Koyeb | H200 | NVIDIA H200 | 141 GB | unknown | Always-on | USD 3/hour | yes | unknown | unknown | stale |
| Koyeb | H200 | NVIDIA H200 | 141 GB | unknown | Scale-to-Zero | USD 3/hour | yes | unknown | unknown | stale |
| Koyeb | 2x A100 SXM | A100 | 80 GB | unknown | Always-on | USD 4.3/hour | yes | unknown | unknown | stale |
| Koyeb | 2x A100 SXM | A100 | 80 GB | unknown | Scale-to-Zero | USD 4.3/hour | yes | unknown | unknown | stale |
| Koyeb | RTX-A6000 | A6000 | 48 GB | unknown | Always-on | USD 0.75/hour | yes | unknown | unknown | stale |
| Koyeb | RTX-A6000 | A6000 | 48 GB | unknown | Scale-to-Zero | USD 0.75/hour | yes | unknown | unknown | stale |
| Koyeb | B200 | NVIDIA B200 | 180 GB | unknown | Always-on | USD 5.5/hour | yes | unknown | unknown | stale |
| Koyeb | B200 | NVIDIA B200 | 180 GB | unknown | Scale-to-Zero | USD 5.5/hour | yes | unknown | unknown | stale |
| Koyeb | 4x H200 | NVIDIA H200 | 141 GB | unknown | Always-on | USD 12/hour | yes | unknown | unknown | stale |
| Koyeb | 4x H200 | NVIDIA H200 | 141 GB | unknown | Scale-to-Zero | USD 12/hour | yes | unknown | unknown | stale |
| Microsoft Azure | NVIDIA A100 | A100 | 80 GB | unknown | NVIDIA A100 — East US | USD 0.000024/core_second + USD 0.000529/second + USD 0.000003/gib_second | no | yes | yes | stale |
| Microsoft Azure | NVIDIA A100 | A100 | 80 GB | unknown | NVIDIA A100 — Sweden Central | USD 0.000024/core_second + USD 0.000688/second + USD 0.000003/gib_second | no | yes | yes | stale |
| Modal | Nvidia A100, 80 GB | A100 | 80 GB | unknown | Broad region | USD 0.00002/core_second + USD 0.001041/second + USD 0.000003/gib_second | no | unknown | unknown | stale |
| Modal | Nvidia A100, 80 GB | A100 | 80 GB | unknown | Narrow region | USD 0.000023/core_second + USD 0.001215/second + USD 0.000004/gib_second | no | unknown | unknown | stale |
| Modal | Nvidia A100, 80 GB | A100 | 80 GB | unknown | Standard region | USD 0.000013/core_second + USD 0.000694/second + USD 0.000002/gib_second | no | unknown | unknown | stale |

## Evidence and interpretation

### Baseten.co: A10G

- Official pricing source: https://www.baseten.co/pricing/
- Data-governance profile: not yet reviewed
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: A10G
- GPU memory per device: 24 GB
- GPU memory total: unknown
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:01:37.097Z
- Warnings: none

### Baseten.co: L4

- Official pricing source: https://www.baseten.co/pricing/
- Data-governance profile: not yet reviewed
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: L4
- GPU memory per device: 24 GB
- GPU memory total: unknown
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:01:37.097Z
- Warnings: none

### Baseten.co: B200

- Official pricing source: https://www.baseten.co/pricing/
- Data-governance profile: not yet reviewed
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: NVIDIA B200
- GPU memory per device: 180 GB
- GPU memory total: unknown
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:01:37.097Z
- Warnings: none

### Baseten.co: H100 MIG

- Official pricing source: https://www.baseten.co/pricing/
- Data-governance profile: not yet reviewed
- Operational profile: not yet reviewed
- Hardware selection: opaque_provider_tier
- Canonical GPU options: none guaranteed
- GPU memory per device: 40 GB
- GPU memory total: unknown
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:01:37.097Z
- Warnings: none

### Baseten.co: H100

- Official pricing source: https://www.baseten.co/pricing/
- Data-governance profile: not yet reviewed
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: H100
- GPU memory per device: 80 GB
- GPU memory total: unknown
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:01:37.097Z
- Warnings: none

### Baseten.co: A100

- Official pricing source: https://www.baseten.co/pricing/
- Data-governance profile: not yet reviewed
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: A100
- GPU memory per device: 80 GB
- GPU memory total: unknown
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:01:37.097Z
- Warnings: none

### Beam.cloud: B200

- Official pricing source: https://www.beam.cloud/pricing
- Data-governance profile: not yet reviewed
- Operational profile: [Beam Serverless](https://cloudgpuprices.com/api/v1/operational-profiles?provider=beam&product=serverless) (verified 2026-07-18)
- Hardware selection: fixed
- Canonical GPU options: NVIDIA B200
- GPU memory per device: 180 GB
- GPU memory total: 180 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:01:41.526Z
- Warnings: unpriced_custom_plan: The Serverless committed-spend option is presented as “With committed spend Book a call →” without a published amount; it was omitted.

### Beam.cloud: RTX 5090

- Official pricing source: https://www.beam.cloud/pricing
- Data-governance profile: not yet reviewed
- Operational profile: [Beam Serverless](https://cloudgpuprices.com/api/v1/operational-profiles?provider=beam&product=serverless) (verified 2026-07-18)
- Hardware selection: fixed
- Canonical GPU options: NVIDIA GeForce RTX 5090
- GPU memory per device: 32 GB
- GPU memory total: 32 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:01:41.526Z
- Warnings: unpriced_custom_plan: The Serverless committed-spend option is presented as “With committed spend Book a call →” without a published amount; it was omitted.

### Beam.cloud: L40S

- Official pricing source: https://www.beam.cloud/pricing
- Data-governance profile: not yet reviewed
- Operational profile: [Beam Serverless](https://cloudgpuprices.com/api/v1/operational-profiles?provider=beam&product=serverless) (verified 2026-07-18)
- Hardware selection: fixed
- Canonical GPU options: L40S
- GPU memory per device: 48 GB
- GPU memory total: 48 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:01:41.526Z
- Warnings: unpriced_custom_plan: The Serverless committed-spend option is presented as “With committed spend Book a call →” without a published amount; it was omitted.

### Beam.cloud: A100 80GB

- Official pricing source: https://www.beam.cloud/pricing
- Data-governance profile: not yet reviewed
- Operational profile: [Beam Serverless](https://cloudgpuprices.com/api/v1/operational-profiles?provider=beam&product=serverless) (verified 2026-07-18)
- Hardware selection: fixed
- Canonical GPU options: A100
- GPU memory per device: 80 GB
- GPU memory total: 80 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:01:41.526Z
- Warnings: unpriced_custom_plan: The Serverless committed-spend option is presented as “With committed spend Book a call →” without a published amount; it was omitted.

### Beam.cloud: H100

- Official pricing source: https://www.beam.cloud/pricing
- Data-governance profile: not yet reviewed
- Operational profile: [Beam Serverless](https://cloudgpuprices.com/api/v1/operational-profiles?provider=beam&product=serverless) (verified 2026-07-18)
- Hardware selection: fixed
- Canonical GPU options: H100
- GPU memory per device: 80 GB
- GPU memory total: 80 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:01:41.526Z
- Warnings: unpriced_custom_plan: The Serverless committed-spend option is presented as “With committed spend Book a call →” without a published amount; it was omitted.

### Beam.cloud: RTX PRO 6000

- Official pricing source: https://www.beam.cloud/pricing
- Data-governance profile: not yet reviewed
- Operational profile: [Beam Serverless](https://cloudgpuprices.com/api/v1/operational-profiles?provider=beam&product=serverless) (verified 2026-07-18)
- Hardware selection: fixed
- Canonical GPU options: NVIDIA RTX PRO 6000 Blackwell
- GPU memory per device: 96 GB
- GPU memory total: 96 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:01:41.526Z
- Warnings: unpriced_custom_plan: The Serverless committed-spend option is presented as “With committed spend Book a call →” without a published amount; it was omitted.

### Beam.cloud: RTX 4090

- Official pricing source: https://www.beam.cloud/pricing
- Data-governance profile: not yet reviewed
- Operational profile: [Beam Serverless](https://cloudgpuprices.com/api/v1/operational-profiles?provider=beam&product=serverless) (verified 2026-07-18)
- Hardware selection: fixed
- Canonical GPU options: RTX 4090
- GPU memory per device: 24 GB
- GPU memory total: 24 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:01:41.526Z
- Warnings: unpriced_custom_plan: The Serverless committed-spend option is presented as “With committed spend Book a call →” without a published amount; it was omitted.

### Beam.cloud: H200

- Official pricing source: https://www.beam.cloud/pricing
- Data-governance profile: not yet reviewed
- Operational profile: [Beam Serverless](https://cloudgpuprices.com/api/v1/operational-profiles?provider=beam&product=serverless) (verified 2026-07-18)
- Hardware selection: fixed
- Canonical GPU options: NVIDIA H200
- GPU memory per device: 141 GB
- GPU memory total: 141 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:01:41.526Z
- Warnings: unpriced_custom_plan: The Serverless committed-spend option is presented as “With committed spend Book a call →” without a published amount; it was omitted.

### Beam.cloud: A6000

- Official pricing source: https://www.beam.cloud/pricing
- Data-governance profile: not yet reviewed
- Operational profile: [Beam Serverless](https://cloudgpuprices.com/api/v1/operational-profiles?provider=beam&product=serverless) (verified 2026-07-18)
- Hardware selection: fixed
- Canonical GPU options: A6000
- GPU memory per device: 48 GB
- GPU memory total: 48 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:01:41.526Z
- Warnings: unpriced_custom_plan: The Serverless committed-spend option is presented as “With committed spend Book a call →” without a published amount; it was omitted.

### Cerebrium: A100 (40GB)

- Official pricing source: https://cerebrium.ai/pricing
- Data-governance profile: [Cerebrium Serverless](https://cloudgpuprices.com/api/v1/governance?provider=cerebrium&product=serverless) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: A100
- GPU memory per device: 40 GB
- GPU memory total: 40 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: yes / yes / unknown
- Reviewed variant regions: Stockholm (eu-north-1, SE) [evidence](https://cerebrium.ai/docs/deployments/multi-region-deployment)
- Verified: 2026-07-20T09:05:10.703Z
- Warnings: none

### Cerebrium: A100 (80GB)

- Official pricing source: https://cerebrium.ai/pricing
- Data-governance profile: [Cerebrium Serverless](https://cloudgpuprices.com/api/v1/governance?provider=cerebrium&product=serverless) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: A100
- GPU memory per device: 80 GB
- GPU memory total: 80 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: yes / yes / unknown
- Reviewed variant regions: Stockholm (eu-north-1, SE) [evidence](https://cerebrium.ai/docs/deployments/multi-region-deployment)
- Verified: 2026-07-20T09:05:10.703Z
- Warnings: none

### Fal.ai: B200

- Official pricing source: https://fal.ai/pricing
- Data-governance profile: not yet reviewed
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: NVIDIA B200
- GPU memory per device: 180 GB
- GPU memory total: 180 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-07-20T09:05:21.291Z
- Warnings: unpriced_custom_plan: The Serverless & Compute Pricing table publishes As low as amounts, but those discounted prices require sales contact and are omitted.

### Fal.ai: B300

- Official pricing source: https://fal.ai/pricing
- Data-governance profile: not yet reviewed
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: NVIDIA B300
- GPU memory per device: 288 GB
- GPU memory total: 288 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-07-20T09:05:21.291Z
- Warnings: unpriced_custom_plan: The Serverless & Compute Pricing table publishes As low as amounts, but those discounted prices require sales contact and are omitted.

### Fal.ai: H100

- Official pricing source: https://fal.ai/pricing
- Data-governance profile: not yet reviewed
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: H100
- GPU memory per device: 80 GB
- GPU memory total: 80 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-07-20T09:05:21.291Z
- Warnings: unpriced_custom_plan: The Serverless & Compute Pricing table publishes As low as amounts, but those discounted prices require sales contact and are omitted.

### Fal.ai: RTX PRO 6000

- Official pricing source: https://fal.ai/pricing
- Data-governance profile: not yet reviewed
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: NVIDIA RTX PRO 6000 Blackwell
- GPU memory per device: 96 GB
- GPU memory total: 96 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-07-20T09:05:21.291Z
- Warnings: unpriced_custom_plan: The Serverless & Compute Pricing table publishes As low as amounts, but those discounted prices require sales contact and are omitted.

### Fal.ai: H200

- Official pricing source: https://fal.ai/pricing
- Data-governance profile: not yet reviewed
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: NVIDIA H200
- GPU memory per device: 141 GB
- GPU memory total: 141 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-07-20T09:05:21.291Z
- Warnings: unpriced_custom_plan: The Serverless & Compute Pricing table publishes As low as amounts, but those discounted prices require sales contact and are omitted.

### Google Cloud: NVIDIA-L4

- Official pricing source: https://cloud.google.com/run/pricing
- Data-governance profile: [Google Cloud Run GPU](https://cloudgpuprices.com/api/v1/governance?provider=google-cloud&product=cloud-run-gpu) (verified 2026-07-18)
- Operational profile: [Google Cloud Run GPU](https://cloudgpuprices.com/api/v1/operational-profiles?provider=google-cloud&product=cloud-run-gpu) (verified 2026-07-18)
- Hardware selection: fixed
- Canonical GPU options: L4
- GPU memory per device: 24 GB
- GPU memory total: 24 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: yes / yes / unknown
- Reviewed variant regions: St. Ghislain, Belgium (europe-west1, BE) [evidence](https://cloud.google.com/run/docs/locations); Eemshaven, Netherlands (europe-west4, NL) [evidence](https://cloud.google.com/compute/docs/regions-zones/gpu-regions-zones)
- Verified: 2026-08-01T00:02:37.037Z
- Warnings: none

### Google Cloud: NVIDIA RTX Pro 6000

- Official pricing source: https://cloud.google.com/run/pricing
- Data-governance profile: [Google Cloud Run GPU](https://cloudgpuprices.com/api/v1/governance?provider=google-cloud&product=cloud-run-gpu) (verified 2026-07-18)
- Operational profile: [Google Cloud Run GPU](https://cloudgpuprices.com/api/v1/operational-profiles?provider=google-cloud&product=cloud-run-gpu) (verified 2026-07-18)
- Hardware selection: fixed
- Canonical GPU options: NVIDIA RTX PRO 6000 Blackwell
- GPU memory per device: 96 GB
- GPU memory total: 96 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: yes / yes / unknown
- Reviewed variant regions: Eemshaven, Netherlands (europe-west4, NL) [evidence](https://cloud.google.com/compute/docs/regions-zones/gpu-regions-zones)
- Verified: 2026-08-01T00:02:37.037Z
- Warnings: none

### Inferless: Nvidia A10 Dedicated

- Official pricing source: https://www.inferless.com/pricing
- Data-governance profile: not yet reviewed
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: A10
- GPU memory per device: 24 GB
- GPU memory total: 24 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:15.760Z
- Warnings: unpriced_custom_plan: Enterprise pricing is listed as discounted or custom and has no published price; omitted from the catalog.

### Inferless: Nvidia A100 Dedicated

- Official pricing source: https://www.inferless.com/pricing
- Data-governance profile: not yet reviewed
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: A100
- GPU memory per device: 80 GB
- GPU memory total: 80 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:15.760Z
- Warnings: unpriced_custom_plan: Enterprise pricing is listed as discounted or custom and has no published price; omitted from the catalog.

### Inferless: Nvidia A100 Shared

- Official pricing source: https://www.inferless.com/pricing
- Data-governance profile: not yet reviewed
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: A100
- GPU memory per device: 40 GB
- GPU memory total: 40 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:15.760Z
- Warnings: unpriced_custom_plan: Enterprise pricing is listed as discounted or custom and has no published price; omitted from the catalog.

### Koyeb: L40S

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: L40S
- GPU memory per device: 48 GB
- GPU memory total: 48 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Koyeb: A100

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: A100
- GPU memory per device: 80 GB
- GPU memory total: 80 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Koyeb: 2x H200

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: NVIDIA H200
- GPU memory per device: 141 GB
- GPU memory total: 282 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Koyeb: 8x A100

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: A100
- GPU memory per device: 80 GB
- GPU memory total: 640 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Koyeb: 4x H100

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: H100
- GPU memory per device: 80 GB
- GPU memory total: 320 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Koyeb: L4

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: L4
- GPU memory per device: 24 GB
- GPU memory total: 24 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Koyeb: 8x H100

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: H100
- GPU memory per device: 80 GB
- GPU memory total: 640 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Koyeb: A100 SXM

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: A100
- GPU memory per device: 80 GB
- GPU memory total: 80 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Koyeb: RTX-PRO-6000

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: NVIDIA RTX PRO 6000 Blackwell
- GPU memory per device: 96 GB
- GPU memory total: 96 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Koyeb: 4x A100 SXM

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: A100
- GPU memory per device: 80 GB
- GPU memory total: 320 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Koyeb: 4x A100

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: A100
- GPU memory per device: 80 GB
- GPU memory total: 320 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Koyeb: 8x H200

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: NVIDIA H200
- GPU memory per device: 141 GB
- GPU memory total: 1128 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Koyeb: H100

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: H100
- GPU memory per device: 80 GB
- GPU memory total: 80 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Koyeb: 2x H100

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: H100
- GPU memory per device: 80 GB
- GPU memory total: 160 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Koyeb: 8x A100 SXM

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: A100
- GPU memory per device: 80 GB
- GPU memory total: 640 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Koyeb: 2x A100

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: A100
- GPU memory per device: 80 GB
- GPU memory total: 160 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Koyeb: H200

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: NVIDIA H200
- GPU memory per device: 141 GB
- GPU memory total: 141 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Koyeb: 2x A100 SXM

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: A100
- GPU memory per device: 80 GB
- GPU memory total: 160 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Koyeb: RTX-A6000

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: A6000
- GPU memory per device: 48 GB
- GPU memory total: 48 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Koyeb: B200

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: NVIDIA B200
- GPU memory per device: 180 GB
- GPU memory total: 180 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Koyeb: 4x H200

- Official pricing source: https://www.koyeb.com/pricing
- Data-governance profile: [Koyeb Serverless Compute](https://cloudgpuprices.com/api/v1/governance?provider=koyeb&product=serverless-compute) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: NVIDIA H200
- GPU memory per device: 141 GB
- GPU memory total: 564 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / unknown
- Reviewed variant regions: none
- Verified: 2026-08-01T00:03:20.927Z
- Warnings: unpriced_custom_plan: Long-term reservations offer up to 50% off, but exact public commitment prices are not published.

### Microsoft Azure: NVIDIA A100

- Official pricing source: https://azure.microsoft.com/en-us/pricing/details/container-apps/
- Data-governance profile: [Azure Container Apps Serverless GPU](https://cloudgpuprices.com/api/v1/governance?provider=azure&product=container-apps-serverless-gpu) (verified 2026-07-18)
- Operational profile: not yet reviewed
- Hardware selection: fixed
- Canonical GPU options: A100
- GPU memory per device: 80 GB
- GPU memory total: 80 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: yes / yes / unknown
- Reviewed variant regions: Italy North (italynorth, IT) [evidence](https://learn.microsoft.com/en-us/azure/reliability/regions-list); Sweden Central (swedencentral, SE) [evidence](https://learn.microsoft.com/en-us/azure/reliability/regions-list)
- Verified: 2026-07-20T09:05:02.368Z
- Warnings: none

### Modal: Nvidia A100, 80 GB

- Official pricing source: https://modal.com/pricing
- Data-governance profile: [Modal Serverless](https://cloudgpuprices.com/api/v1/governance?provider=modal&product=serverless) (verified 2026-07-18)
- Operational profile: [Modal Serverless Functions](https://cloudgpuprices.com/api/v1/operational-profiles?provider=modal&product=serverless) (verified 2026-07-18)
- Hardware selection: fixed
- Canonical GPU options: A100
- GPU memory per device: 80 GB
- GPU memory total: 80 GB
- Included local storage: unknown
- EU compute / constrained / EU-only: unknown / unknown / no
- Reviewed variant regions: none
- Verified: 2026-07-19T11:33:02.043Z
- Warnings: none

## Next page

[Continue with the next page](https://cloudgpuprices.com/pricing.md?category=serverless_gpu&min_gpu_memory_gb=24&cursor=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)