TensorDock floor

tensordock.com · Compute · 2 records · JSON

TensorDock Computepaidemail?card?Drequest-rate
TensorDock floor ComputepaidcardcardR

TensorDock Compute

paidemail?card?Drequest-rate

www.tensordock.com/ · GPU cloud · compute/tensordock · JSON · compare

Free tier	no free tier: GPU servers from $0.12 an hour for a consumer card to $2.25 for an H100, with a $5 minimum to start and root access unit none; metered on none
Price	none stated
Per-object cap	not stated
Retention	not stated
Egress	not stated
Identity	email (placeholder) not published
Limits	request-rate from $0.12 an hour for a consumer GPU to $2.25 for an H100 SXM5 source write-gated a $5 minimum balance is needed to launch a server source per-object-cap 45 GPU models are offered, with root access and dedicated cards source excluded as gpu-only
Binding	request-rate
Interface	unknown
Tooling	not stated
Composes with	not stated
Reliability	no reliability event recorded
Evidence rung	documented the site was read on 2026-10-22 and states the GPU ladder, the $5 starting point and root access. There is no free tier, so the verdict moves to paid-only
Checked	2026-10-23 (live); stale on 2026-11-22

Evidence (2)

date	source	what it proved
2026-10-02	www.tensordock.com/	not published
2026-10-22	www.tensordock.com/	'H100s from $2.25/hr. No quotas, no commitments.' with 'Start with just $5 and launch a server in 30 seconds', 'consumer GPUs from $0.12/hr', '45 available GPU models' and 'Root access and dedicated GPUs', quoted

More fields (14)

{
 "talaria_m10h_usd": null,
 "talaria_m24h_usd": null,
 "nemo_m10h_usd": null,
 "keep_rate": null,
 "keep_rate_basis": "no field in the source",
 "billing_model": "unclassified",
 "shape": {
  "vcpu": null,
  "ram": null,
  "disk": null,
  "gpu_class": "RTX 4090 to HGX H100 SXM5",
  "gpu_count": 1,
  "note": "45 GPU models, from consumer cards at $0.12 an hour to an HGX H100 SXM5 at $2.25"
 },
 "session_cap": {
  "wall_clock_min": null,
  "idle_reap_s": null,
  "restart_required": false,
  "note": "a server launches in 30 seconds and bills while it exists; the page states no reap"
 },
 "persistent_storage": null,
 "network_out": null,
 "price_note": null,
 "free_tier_kind": "none",
 "ps": {
  "gib": null,
  "price": "storage is part of the server",
  "durability": "the server's own disk",
  "on_stop": true,
  "on_destroy": false,
  "retention": "while the server exists"
 },
 "net": {
  "out": true,
  "in": true,
  "ports": "root access with full OS control",
  "static_ip": true,
  "ipv6": null,
  "allowance": null
 }
}

Unknown or withheld values (15)

$.chaining not-stated
$.egress.metered egress-unknown
$.egress.terms not-stated
$.evidence[0].proved process
$.extras.price_note process
$.identity_cost.detail defaulted
$.identity_cost.kind identity-defaulted
$.interface interface-unknown
$.interface_detail withheld
$.per_object_cap.notes process
$.per_object_cap.stated not-stated
$.reliability_note none-observed
$.retention.notes not-extracted
$.retention.stated not-stated
$.tooling not-stated

TensorDock floor Compute

paidcardcardR

www.tensordock.com/ · GPU cloud · compute/tensordock-floor · JSON · compare

Free tier	The vendor's home page quotes 'consumer GPUs from $0.12/hr' and 'H100s from $2.25/hr' across '45 available GPU models', with a USD 5 minimum to start. An independent tracker on the same day prices the cheapest TensorDock cards at USD 0.10 for an RTX A4000 16GB, USD 0.17 for a Tesla V100 16GB, USD 0.20 for an RTX 3090 24GB, USD 0.35 for an RTX 4090 24GB and USD 0.75 for an RTX 6000 Ada, which is below the vendor's own published floor. unit GiB; metered on wall-clock; none
Price	none stated
Per-object cap	not applicable
Retention	not applicable
Egress	not stated
Identity	card Unchanged from the existing record: a USD 5 minimum balance is the entry charge.
Binding	none The binding constraint is the host's, not the platform's: prices are set by third-party hosts on a marketplace, so the card exists at the rate but the rate moves with supply.
Interface	unknown inferred from the tooling the provider documents; not probed
Tooling	not stated
Composes with	A USD 0.10 16 GB hour is the smallest possible entry, and with a USD 5 minimum deposit it buys 50 card-hours, which is a complete small fine-tune for the smallest possible outlay.
Reliability	no reliability event recorded
Notes	Corrects the existing tensordock.json record, which says only 'GPU servers from USD 0.12 an hour for a consumer card'. The two sources disagree on the floor, 0.12 against 0.10, and I did not resolve which configuration each names: the vendor figure is a rounded marketing line and the tracker figure is per card type. The honest statement is that the floor is between USD 0.10 and USD 0.12. The vendor page also claims 45 models, which is more than any other single host.
Evidence rung	reported
Checked	2026-10-22 (live); stale on 2026-11-05

Evidence (3)

date	source	what it proved
2026-10-22	www.tensordock.com/	quoted: 'consumer GPUs from $0.12/hr'; 'H100s from $2.25/hr. No quotas, no commitments'; '45 available GPU models'; 'Start with just $5'
2026-10-22	gpurentalprices.com/rtx-a4000	quoted: RTX A4000 on-demand from USD 0.10 at TensorDock across 3 providers, below the vendor's own USD 0.12 claim
2026-10-22	gpurentalprices.com/providers/tensordock	the tracker reads TensorDock as the on-demand floor for RTX 4090 at USD 0.35, RTX 3090 at USD 0.20, V100 at USD 0.17 and RTX 6000 Ada at USD 0.75

More fields (1)

{
 "wave2": {
  "class": "gpu-marketplace",
  "free_tier_kind": "none",
  "shape": {
   "vcpu": null,
   "ram_gib": null,
   "gpu_class": "RTX A4000",
   "gpu_count": 1,
   "vram_gib": 16
  }
 }
}

Unknown or withheld values (9)

$.egress.metered egress-unknown
$.egress.terms not-stated
$.interface interface-unknown
$.per_object_cap.notes process
$.per_object_cap.stated not-applicable
$.reliability_note none-observed
$.retention.notes process
$.retention.stated not-applicable
$.tooling not-stated