Book a two-hour remote session on a live 8-GPU server, reserve units from the first production allocation, or join early access for Nova 8 Training.
Book a two-hour slotPick a time. We confirm within one working day.
Log in remotelyA fresh, isolated environment on a live Nova 8. Bring your model and data, under NDA if needed.
See your own tokens per secondLive throughput, power and temperature on your workload. Everything is wiped when you finish.
Pre-order with confidenceIf the numbers hold on your workload, reserve units from the first production allocation.
Each session runs in its own isolated environment that's created for you and wiped when the session ends. We can sign an NDA before you upload anything.
Llama 3.3 70B is pre-loaded and benchmarked. You can bring your own open-weight model in standard formats; tell us in advance and we'll have the right inference stack ready.
Tokens per second for your model and prompts, plus live GPU power and temperature, so you can compare with what you run today.
No. A pre-order reserves an allocation. No payment is taken until a Renda specialist has confirmed configuration, pricing and lead time with you.
Nova 8 Training is in development. Early-access members hear first about engineering samples and test slots.