Dedicated model · Available as managed deployment
OpenAI's open-weight reasoning models — 20B and 120B mixture-of-experts — built for agentic work, tool use and long chains of reasoning, under Apache-2.0. Validated on AxForge hardware and deployed on a dedicated DGX Spark for your traffic only — an OpenAI-compatible endpoint on hardware only you use, operated by AxForge in the EU.
Why AxForge
| OpenAI, open weights | The first open-weight models from OpenAI: two sizes, both released under Apache-2.0, with configurable reasoning effort and native tool calling. |
|---|---|
| 20B on one machine | gpt-oss-20b is sized for a single accelerator; on a dedicated DGX Spark it runs as your own OpenAI-compatible endpoint. The 120B build is a multi-GPU deployment. |
| Agents, not just chat | Structured outputs, function calling and a reasoning mode you can dial down for latency — the pieces an agent workflow needs, on hardware only you use. |
Specifications
| Model | gpt-oss — openai |
|---|---|
| Modalities | Text |
| Sizes | 20.9B, 116.8B |
| Licence | Open weights — apache-2.0; commercial use permitted |
| Hardware | NVIDIA DGX Spark (GB10, 128 GB unified memory) — owned and operated by AxForge |
| Rental term | Hour, week, month or year |
| Hardware pricing | €0.69/hour on demand · €0.66/hour by the week · €0.62/hour by the month · €0.55/hour by the year, excl. VAT |
| Managed service | Quoted per deployment |
| Region | Málaga, Spain (eu-es-1) |
Full details, benchmarks and FAQ on the gpt-oss page. Prices exclude VAT.
How it works
| 1 | Request deployment — describe your traffic, context needs and rental term. |
|---|---|
| 2 | You receive the configuration, hardware rental and managed-service price in writing before anything is billed. |
| 3 | AxForge deploys gpt-oss on a dedicated DGX Spark reserved for you. |
| 4 | Point your OpenAI SDK at your own endpoint with the model name you receive. |
| 5 | Adjust the term — hour, week, month or year — as your workload settles. |
Request deployment or sign in to start.
FAQ
Not on the serverless API — it is available as a managed deployment: validated on AxForge hardware and deployed on a dedicated DGX Spark for your traffic only. The serverless API serves Qwen3.8 27B.
gpt-oss-20b for a single dedicated machine and low latency; gpt-oss-120b when quality on hard reasoning matters more than cost — it needs a multi-GPU system, which AxForge scopes with you.
Yes — function calling and structured outputs are part of the model, and they work through the OpenAI-compatible endpoint on your machine.
AxForge publishes only numbers it measures itself, and has not benchmarked this model on its nodes yet. For quality benchmarks, see the official model card.
Hardware by the hour, week, month or year; the managed service is quoted per deployment — both confirmed in writing before anything is billed.