runpod network
runpod.io · Compute, Storage · 2 records · JSON
- RunPod ComputepaidcardcardDrequest-rate
- runpod network volumes Storagepaidemail?card?R
RunPod Compute
paidcardcardDrequest-rate
| Free tier | no free GPU tier is published; the on-demand card list runs from an RTX A5000 at USD 0.16 an hour on Community Cloud to a B300 at USD 8.99 on Secure Cloud, across 21 cards unit none; metered on wall-clock; Community Cloud is cheaper than Secure Cloud for the same card; the page carries a per-hour and a per-second toggle, and 21 on-demand cards were read on 2026-10-24 |
|---|---|
| Price | none stated |
| Per-object cap | not applicable |
| Retention | network storage is billed while it exists; a pod keeps its volume only while attached or persisted |
| Egress | not published as a per-GB line on the page read |
| Identity | card a card is required to launch a pod |
| Limits |
storage from USD 0.05 per GB a month on standard network storage |
| Binding | request-rate billing is per second with no free allowance, so the metered rate is what binds; there is no quota to exhaust |
| Interface | rest, cli REST API, CLI and a web console; SSH into a pod |
| Tooling | runpodctl, the RunPod API, Docker images, network volumes |
| Composes with | a GPU pod is a server with a GPU, so it can hold a model and serve it; the free tier does not exist |
| Reliability | no reliability event recorded |
| Evidence rung | documented |
| Checked | 2026-10-22 (live); stale on 2026-11-21 |
| Parsed numbers |
|
Evidence (3)
| date | source | what it proved |
|---|---|---|
| 2026-10-22 | www.runpod.io/pricing | on-demand B300 at USD 6.94/h on Community Cloud and USD 8.99/h on Secure Cloud; H200 at USD 3.59 and USD 5.29 |
| 2026-10-22 | www.runpod.io/pricing | network storage starts at USD 0.05 per GB a month |
| 2026-10-24 | www.runpod.io/pricing | the full on-demand card list, Community and Secure, quoted from the page's own rate data: RTX A5000 0.16/0.27, RTX 3090 0.22/0.50, RTX A6000 0.33/0.59, RTX 4090 0.34/0.89, A40 0.35/0.59, L4 0.44/0.59, Pro 6000 MIG 24GB 0.50/0.69, L40 0.69/0.82, RTX 5090 0.69/1.19, RTX 6000 Ada 0.74/0.99, L40S 0.79/1.09, Pro 6000 MIG 48GB 1.00/1.29, A100 PCIe 1.19/1.79, A100 SXM 1.39/1.79, RTX Pro 6000 1.69/2.49, H100 PCIe 1.99/2.89, H100 NVL 2.59/3.19, H100 SXM 2.69/3.99, H200 3.59/5.29, B200 5.98/7.99, B300 6.94/8.99. The page carries a per-hour and a per-second toggle, and the B300 and H200 pairs match the 2026-10-22 read |
More fields (12)
{
"shape": {
"vcpu": null,
"ram_gib": null,
"disk_gib": null,
"gpu_class": "B300",
"gpu_count": 1
},
"billing_model": "allocated",
"granularity": {
"min_billed_s": 1,
"rounding_s": 1
},
"session_cap": {
"wall_clock_min": null,
"idle_reap_s": null,
"restart_required": false,
"note": "none; a pod runs until stopped"
},
"keep_rate": {
"value": 0.0,
"basis": "a stopped pod does not bill compute, only storage",
"source": "https://www.runpod.io/pricing"
},
"persistent_storage": {
"included_gib": null,
"price_gib_month": "USD 0.05 and up",
"durability": "network volume",
"survives_stop": true,
"survives_destroy": false,
"max_retention": "while the volume exists"
},
"network_out": {
"allow_outbound": true,
"allow_inbound_listen": true,
"ports": "a TCP port map and a proxy",
"static_ip": null,
"ipv6": null,
"per_session_bytes": null
},
"free_tier_kind": "none",
"m10h_usd": 1.6,
"m24h_usd": 116.8,
"price_arithmetic": "USD 0.16/h is the cheapest card on the page read on 2026-10-24 (RTX A5000, Community Cloud); 10 h = USD 1.60 and 730 h = USD 116.80.",
"burst_x": 1.0
}Unknown or withheld values (6)
$.egress.meteredegress-unknown$.extras.price_arithmeticpartly-held$.noteswithheld$.per_object_cap.notesprocess$.per_object_cap.statednot-applicable$.reliability_notenone-observed
runpod network volumes Storage
paidemail?card?R
| Free tier | Correct architecture, wrong economics: network volumes are $0.07/GB/mo for the first 1 TB ($0.05 beyond), i.e. ~$72/TB/mo, and there is no free allowance. An account at $0 balance has its vo unit none; metered on none |
|---|---|
| Price | none stated |
| Per-object cap | not recorded |
| Retention | not recorded |
| Egress | not recorded |
| Identity | email (placeholder) not published |
| Binding | none no constraint of a named kind was recorded; the verdict rests on the quota sentence rather than on a named limit |
| Interface | unknown |
| Tooling | not recorded |
| Composes with | not recorded |
| Reliability | Correct architecture, wrong economics: network volumes are $0.07/GB/mo for the first 1 TB ($0.05 beyond), i.e. ~$72/TB/mo, and there is no free allowance. An account at $0 balance has its vo |
| Notes | Correct architecture, wrong economics: network volumes are $0.07/GB/mo for the first 1 TB ($0.05 beyond), i.e. ~$72/TB/mo, and there is no free allowance. An account at $0 balance has its vo |
| Evidence rung | reported |
| Checked | 2026-10-21 (inherited); stale on 2026-11-04 |
Evidence (2)
| date | source | what it proved |
|---|---|---|
| 2026-10-21 | docs.runpod.io/storage/network-volumes | Correct architecture, wrong economics: network volumes are $0.07/GB/mo for the first 1 TB ($0.05 beyond), i.e. ~$72/TB/mo, and there is no free allowance. An account at $0 balance has its vo |
| 2026-10-21 | www.runpod.io/pricing | cited as the basis of the rejection: Correct architecture, wrong economics: network volumes are $0.07/GB/mo for the first 1 TB ($0.05 beyond), i.e. ~$72/TB/mo, and there is no free allowance. An account at $0 balance has its vo |
More fields (1)
{
"dead_end": {
"failure_class": "paid",
"why": "Correct architecture, wrong economics: network volumes are $0.07/GB/mo for the first 1 TB ($0.05 beyond), i.e. ~$72/TB/mo, and there is no free allowance. An account at $0 balance has its vo",
"retry_when": null,
"first_recorded": "2026-10-21"
}
}Unknown or withheld values (15)
$.chainingnot-recorded$.confidence_notewithheld$.egress.termsnot-recorded$.extras.dead_end.retry_whenprocess$.free_tier.notesprocess$.identity_cost.detaildefaulted$.identity_cost.kindidentity-defaulted$.interfaceinterface-unknown$.interface_detailwithheld$.limits_noteprocess$.per_object_cap.notesnot-recorded$.per_object_cap.statednot-recorded$.retention.notesprocess$.retention.statednot-recorded$.toolingnot-recorded