Compute built for AI
A network of GPU hosts supplies the model execution. You can use that capacity through Inferum without deploying model servers or managing GPU infrastructure.
Gonka brings together GPU providers to run AI models. Inferum makes that compute available to your applications through one familiar API.
Gonka is a decentralized network where independent hosts contribute GPU capacity and run AI models. Inference is the work a model does to turn your input into a response, such as an answer, a piece of code, or a summary.
The models run on the hosts’ servers. The network protocol coordinates work, verification, and rewards for the providers supplying that compute.
Read about GonkaGonka gives us access to distributed GPU capacity for open models. That lets us focus on our goal: affordable inference that is straightforward to use in applications, coding tools, and agents.
A network of GPU hosts supplies the model execution. You can use that capacity through Inferum without deploying model servers or managing GPU infrastructure.
Choose a model from our catalogue for text, code, or reasoning. The same Inferum endpoint and API key work across the models we offer.
Public explorers show network participants, inference activity, and transactions. Gonka.gg gives you a way to look at the network behind the service.
Use an OpenAI-compatible client with your Inferum endpoint and key. We handle the connection to the network.
Choose a model and send your messages to the Inferum API using your existing client or SDK.
We check your API key and available balance, then route the request to Gonka through OpenBroker.
Gonka hosts run the model. Inferum delivers its response, including streaming when requested, and records your usage.
Inferum brings API keys, streaming responses, usage history, prepaid billing, and support into one service. You can use Gonka models without operating a network node or managing GNK yourself.
Your account balance and request history stay in your Inferum dashboard. Our model catalogue and pricing page show what is available through Inferum and what it costs.
Browse network activity, inference statistics, participants, and transactions on gonka.gg.