RunPod vs Thunder Compute vs Modal for H100
Every H100 option these three vendors list publicly, priced for 100 GPU hours and a full month. Cheapest among tracked providers first.
| # | Vendor | H100 option | Rate | 100 h/month | Full month (730 h) | Vendor quote |
|---|---|---|---|---|---|---|
| 1 | RunPod | H100 PCIe · Community Cloud | $1.99/hr | $199.00 | $1,452.70 | “H100 PCIe $1.99/hr” |
| 2 | RunPod | H100 SXM · Community Cloud | $2.69/hr | $269.00 | $1,963.70 | “H100 SXM $2.69/hr” |
| 3 | RunPod | H100 PCIe · Secure Cloud | $2.89/hr | $289.00 | $2,109.70 | “H100 PCIe 80GB $2.89/hr” |
| 4 | Thunder Compute | H100 SXM · on-demand | $3.2/hr | $320.00 | $2,336.00 | “H100 SXM 80GB $3.20/hr” |
| 5 | Modal | H100 SXM · Serverless (per second) | $3.9492/hr | $394.92 | $2,882.92 | “Nvidia H100 SXM5 $0.001097/sec” |
| 6 | RunPod | H100 SXM · Secure Cloud | $3.99/hr | $399.00 | $2,912.70 | “H100 SXM 80GB $3.99/hr” |
How they bill
| Vendor | Lowest-priced H100 (tracked) | Billing and extras |
|---|---|---|
| RunPod | H100 PCIe (Community Cloud) $1.99/hr | Pods billed per hour (per-second view available); storage extra, network volumes from $0.05/GB/mo. |
| Thunder Compute | H100 SXM (on-demand) $3.2/hr | Instances billed per minute while running; first 100 GB disk per GPU included; no egress fees. |
| Modal | H100 SXM (Serverless (per second)) $3.9492/hr | Serverless, billed per second while running; listed rates are preemptible (non-preemptible 3x); $30/month free compute on Starter. |
Read before comparing
Rankings are the cheapest among tracked providers on our check date, not across the whole market. Vendors we don't track may cost less.
GPU-hours vs wall-clock time
Prices are per GPU-hour: one GPU for one hour. Four GPUs for 10 wall-clock hours is 40 GPU-hours. Per-second billing (serverless) charges only while your code runs; hourly pods bill from start to stop, idle time included; monthly servers bill a flat price whatever you use. A faster GPU can finish the same job in fewer GPU-hours, so the lowest hourly rate isn't always the lowest cost per job.
Hardware differences
The same GPU name can mean different hardware. SXM modules link GPUs over NVLink and run at higher power limits than PCIe cards, so an H100 SXM is usually faster than an H100 PCIe, especially for multi-GPU training. Interconnect between GPUs and nodes (NVLink, InfiniBand, plain Ethernet) is rarely listed on pricing pages; treat it as unknown unless the vendor states it. VRAM decides which models fit at all. Hardware as listed: H100 PCIe 80 GB (PCIe); H100 SXM 80 GB (SXM).
Egress and storage extras
Totals on this page exclude these. Each cell cites the vendor page it came from; unknown means the vendor doesn't publish it there.
| Vendor | Egress | Storage | Interruption / billing terms |
|---|---|---|---|
| Modal | Network egress $0.04/GiB; Starter plan includes 1 TiB/month free. (source, 2026-10-27) | Volumes $0.09/GiB/mo (includes 1 TiB/mo free). (source, 2026-10-27) | Functions are preemptible by default; non-preemptible execution is 3x the base price. CPU and memory are billed separately. (source, 2026-10-27) |
| RunPod | unknown: Not published on the pricing page. (checked 2026-10-27) | Container disk $0.10/GB/mo; volume disk $0.10/GB/mo running, $0.20/GB/mo idle; network storage $0.07/GB/mo under 1 TB, $0.05/GB/mo over 1 TB. (source, 2026-10-27) | Listed rates are on-demand Pod prices (Secure Cloud and Community Cloud). (source, 2026-10-27) |
| Thunder Compute | No egress fees ($0.00/GB). (source, 2026-10-27) | Instance disk: first 100 GB per GPU included, then $0.219/GB/mo while running; snapshots $0.05/GB/mo. (source, 2026-10-27) | Instances and sandboxes are never preempted. Instances (including extra vCPUs and disk) are billed per minute while running. (source, 2026-10-27) |
Interruptible vs on-demand
Interruptible (preemptible or spot) capacity can be stopped by the vendor mid-job and is usually cheaper; on-demand runs until you stop it. Interruptible (preemptible) rates listed here: Modal H100 SXM. All other rates are on-demand: the vendor doesn't publish preemption for them.
FAQ
Which of RunPod, Thunder Compute and Modal is cheapest among tracked providers for an H100?
On 2026-10-27, the lowest listed H100 rate of the three was RunPod H100 PCIe (Community Cloud) at $1.99/hr, or $199.00 for 100 GPU hours.
What is the cheapest tracked H100 outside a community marketplace?
RunPod H100 PCIe (Secure Cloud) at $2.89/hr.
When does Modal cost less than an hourly H100?
When your GPU sits idle a lot. Modal charges per second only while code runs, so a spiky inference workload can cost less even at a higher listed rate. For a GPU busy most of the hour, the lower hourly rate wins.
What isn't included?
Storage, extra CPU/RAM, egress and taxes (see the cited extras above). Confirm on each vendor page before you rent.
More
- All verified GPU rental prices
- H100: cheapest among tracked providers
- H200: cheapest among tracked providers
- A100: cheapest among tracked providers
- L40S: cheapest among tracked providers
- RTX 4090: cheapest among tracked providers
- RTX 5090: cheapest among tracked providers
- RTX A6000: cheapest among tracked providers
- GPU price per GB of VRAM
- RunPod Community vs Secure Cloud
- GPU cost calculator
- GPU & compute deals
Using an AI agent? Add Lantern Works to your agent (free MCP server, read-only).