Models
Every model below runs on hardware we own. Prices are per 1M tokens, in credits (1 credit = $0.01). This page reads from the same catalog billing charges against, so it cannot drift.
Qwen3 Coder 30B A3B
chat-defaultAgentic coding and long-context work. Mixture-of-experts, served on our GPU.
Context64K
Input / 1M10 cr$0.10
Output / 1M30 cr$0.30
Qwen 3.6
qwen3.6Coding and thinking. 256K context. Served on our own GPU node.
Context256K
Input / 1M40 cr$0.40
Output / 1M120 cr$1.20
More machines are in commissioning, and the catalog grows as they come online. For a specific model or reserved capacity, talk to us. Or request access now; invited accounts start with 100 free credits.