Tool profile · Publisher documentation
TensorDock is a cloud GPU and CPU platform for running AI workloads. According to its supplier's public documentation, you can deploy on-demand GPU servers for training and inference of AI models, choosing from 45 GPU models on KVM virtual machines with root access and an API for server management. A CPU cloud option targets transcoding and batch processing. Published prices start at 0.012 USD per hour for CPU Cloud, 0.12 USD per hour for consumer GPUs, 0.35 USD per hour for an RTX 4090, 1.80 USD per hour for an A100 SXM4 and 2.25 USD per hour for an H100 SXM5, with a 5 USD starting amount. LE LABO IA has not tested TensorDock; this profile relies on the public documentation of its supplier. Unconfirmed prices, quotas, availability and terms should be checked with the supplier.
Tool profile · Publisher documentation
The product points below come from publisher documentation. No common hands-on test is claimed. Documentary profile, sources checked October 7, 2026. LE LABO IA has not tested this product.
Deploy on-demand cloud GPU servers for training AI models, based on the supplier's documented description. Run inference of AI models on rented GPU capacity instead of owning hardware. Pick among 45 GPU models according to the workload, as listed in the supplier's documentation. Use KVM virtual machines with root access for custom software setups and system-level configuration.
Published prices read from the supplier's home page are stated in USD per hour: CPU Cloud from 0.012, consumer GPUs from 0.12, RTX 4090 from 0.35, A100 SXM4 from 1.80 and H100 SXM5 from 2.25, with billing by usage and a starting amount of 5 USD via a funds deposit. These amounts come from a single page and are not confirmed. Treat the hourly rates, the billing model and the starting amount as points to check with the supplier before committing.
No test of the product was performed by LE LABO IA, so behaviour, reliability and support quality remain unknown. The published prices come only from the supplier's home page and are not confirmed; rates and billing terms may change or differ by machine and period. Quotas, territorial availability and terms of use are not covered by the pricing information available and are not confirmed; check them with the supplier.
Before committing, visit the supplier's site and check a small on-demand GPU offer against your own workload. Try a short training or inference run to see how setup, access and billing behave in practice. Also confirm the published hourly rates, any quotas and the availability in your region with the supplier, since these points are not confirmed.
