Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

Contabo GPU VPS is live

contabo_mcontabo_m Member, Patron Provider

Hey LET,

GPU VPS is something we've been asked about for a while. It's live now.

One dedicated NVIDIA RTX PRO 6000 per instance, flat monthly rate. Same price every month, regardless of usage

What you get:

NVIDIA RTX PRO 6000 - 96 GB GDDR7 VRAM, Blackwell, 24,064 CUDA cores

18 vCPUs, 96 GB RAM, 900 GB NVMe

Ubuntu 24.04 with CUDA preinstalled, works from first login

Unlimited traffic (fair use applies)

€999/mo | $1,199/mo | £969/mo - 15% off 12-month, 20% off 24-month

Storage Extension, Private Networking, and 1-Click Apps (Ollama, n8n, OpenClaw, Hermes) all supported.

On the card: 96 GB VRAM is enough to run a 70B model on a single GPU without quantization. FP4 via 5th-gen Tensor Cores handles larger or longer-context workloads without needing to split across instances.

Starting with a limited run in EU and US Central, since GPU capacity is genuinely constrained in a way that CPU and storage aren't.

Questions welcome.

For more details : https://contabo.com/en/gpu-vps
Blog link: https://contabo.com/blog/gpu-vps-is-here-dedicated-nvidia-gpu-power-without-the-hyperscaler-price/

Team Contabo

Comments

  • First

  • x0x0xx0x0x Member

    Couldn't make it into the Offers section, so it becomes "News" :p

  • rpqurpqu Member

    hourly option?

    Thanked by 1barbarza
  • AstroAstro Member

    Cant people get a dedi with those/better specs for lower pricing? Honest question - I have no idea

    Thanked by 2barbarza monsoon73
  • @Astro said:
    Cant people get a dedi with those/better specs for lower pricing? Honest question - I have no idea

    You can 96GB VRAM is a crazy GPU. You can probably find it cheaper on hourly basis + contabo is meh.

  • Unironically why would anyone buy such a card.

    For models on the <70B range, there are so many better options.

    Any model not on GPT-OSS level is not worth paying monthly to have external hosted, and at this cost you could simple pay other AI hosts not sure how it's a seeling point.

    A Mac Mini clustered can perform as good as this card, for 40% less owned, in your house with ample compute left over.

Sign In or Register to comment.