Pricing
Forge time. Nothing more.
No subscriptions. No minimums. Raw GPU compute billed by the minute, starting at $0.50/hr. You pay for the heat, not the furnace.
FLASH
Qwen 2.5 7B
The first fracture. Fast extraction and quick analysis — ideal for straightforward document review, summaries, and data extraction from raw material.
7B parameters, AWQ quantization, 1x 24GB GPU
AIR
Qwen 2.5 32B
Deep reasoning across complex documents. Nuanced analysis, multi-step queries, and balanced performance for most professional workflows requiring sustained heat.
32B parameters, AWQ quantization, 4x 48GB GPU
PRO
Qwen 2.5 72B
Full-precision reasoning. Expert-level document intelligence forged for demanding legal, financial, and medical workloads where nothing can be left unexamined.
72B parameters, AWQ quantization, 4x 48GB GPU
MAX
Qwen 2.5 72B FP16
The deepest vein. Uncompromising analytical depth with full-precision weights and no quantization trade-offs. For those who demand obsidian-grade clarity from every inference.
72B parameters, FP16 full precision, 8x 80GB GPU
Deepest vein
Billing
The exchange
Load the furnace
Top up your prepaid balance using Visa or Mastercard. Minimum top-up is $5.00. No subscriptions, no recurring charges. Your credit sits until you are ready to forge.
Burn by the minute
GPU time is billed per minute at the hourly rate for your chosen tier. A $0.50/hr vault costs less than a penny per minute. You pay only for active compute — nothing smolders in the background.
Watch the flame
Track your balance and usage in real-time from your dashboard. When your balance runs low, the vault auto-pauses to prevent unexpected charges. No surprises. No overruns.
FAQ
Questions
No. There are no subscriptions, no minimums, and no contracts. You add credit and use it at your own pace. Unused balance stays in your account indefinitely.
Your vault will auto-pause when your balance is insufficient to continue running. Your data remains on the instance. Top up your balance to resume, or destroy the vault to release all resources.
Each vault is forged with a specific model tier chosen at creation. To use a different tier, create a new vault. You can run multiple vaults simultaneously on different tiers.
Billing is calculated to the exact minute. If you run a FLASH vault for 23 minutes, you pay for 23 minutes at $0.50/hr, which is approximately $0.19.
No. You only pay for active GPU time. Paused vaults do not incur charges. However, the instance remains allocated and will resume billing when restarted.
We accept Visa and Mastercard. All payments are processed securely with PCI DSS compliance.