Catalog
Initial beta catalog
A curated catalog of open models selected for useful capability, real demand, and efficient inference. Throughput and serving detail for each model are on the benchmarks page.
| Model | Best for | Context | Availability |
|---|---|---|---|
| Qwen3.8-27B | Coding, agents, structured output | 128K | Beta |
Per-token pricing will be published before beta access opens. Prices are set from measured serving cost, so they are fixed when they are quoted.
The catalog rotates. Models are added when they meet our demand, performance and economics requirements, and retired through a published lifecycle process.
We don't try to host everything.
Every model we evaluate has to clear the same bar. A model is added when it earns its place in the catalog, not because it is new.
How models are selected- Real demand from developers running recurring workloads
- Useful capability for the workloads people actually run
- Serving efficiency on commodity hardware
- Licensing we can serve commercially
- Measured, reproducible performance
- Pricing that stays profitable as inference economics move
Under evaluation
Models we are actively assessing. Nothing here is available to call yet, and it is listed so the pipeline is visible rather than implied.
Ornith-1.5-35B-A3B
Coding · Agents
Not yet benchmarked on our serving configuration.
Ternary-Bonsai-2-27B
General · Efficient
Not yet benchmarked on our serving configuration.