Packet.ai
On-demand NVIDIA Blackwell GPU cloud with per-second billing, SSH, CLI, and an OpenAI-compatible inference API
Packet.ai is an on-demand GPU cloud for AI and ML workloads, built by hosted.ai in Europe. It offers NVIDIA B200, H200, and RTX PRO 6000 Blackwell (96GB) GPUs with full root SSH, a web dashboard, CLI, and API access. Instances are on-demand rather than spot, with a published 99.9% uptime SLA and per-second billing, so there are no contracts or preemption.
It also provides a Token Factory API, a managed inference endpoint that is OpenAI-compatible and billed per token, for teams that want model inference without managing GPUs directly.
Packet.ai positions on price, listing rates such as $0.66/hr for an RTX PRO 6000 Blackwell, achieved by scheduling workloads to reduce idle GPU time. Setup runs from signup to SSH in under five minutes, with payment via card, bank transfer, or crypto.
Pricing: Pay-as-you-go
Packet.ai Alternatives
Explore 76 products in the Inference APIs category. View all Packet.ai alternatives.
CheapestInference
Flat-rate unlimited inference on open-weight models, sold in daily 8-hour windows
Lyceum
EU-hosted inference cloud for open-source models, OpenAI-compatible
vLLM
High-throughput LLM inference engine with PagedAttention for efficient GPU memory usage
Work on Packet.ai? Feature it at the top of Inference APIs.
Is your product missing?