What changed
Hugging Face has announced that Baseten is now a supported inference provider on its platform. This addition expands the range of serverless inference options available directly within model pages on the Hub. The integration includes seamless support in client SDKs for both JavaScript and Python environments.
Why it matters for practical AI work
Developers can now access a catalog of frontier models, including specific large language models such as Kimi K3, DeepSeek V4 Flash, and GLM-5.2 through this new provider route. The system routes requests automatically when authenticated via the Hugging Face Hub token, passing costs directly to Baseten without additional markup from the platform.
What a reader can verify
Users can verify model availability by checking the full list of models supported by Baseten provided in the source documentation. Feedback on this feature is being collected via a dedicated discussion space hosted at huggingface.co/spaces/huggingface/HuggingDiscussions.
What remains unknown
It remains unknown when support for additional tasks beyond conversational and text-generation will be rolled out. The source material does not specify the exact timeline for these future expansions or confirm if revenue-sharing agreements have already been established with provider partners.
Source-grounded facts
- Baseten is now a supported inference provider on the Hub.
- Support launches for conversational and text-generation tasks.
What remains unknown
- The exact timeline for rolling out support to additional model types remains unspecified in the source text.
Frequently asked questions
How does the new provider integration work technically?
The integration allows developers to use Baseten-hosted models directly through client SDKs for both JavaScript and Python without extra glue code.
What billing benefits are available to PRO subscribers?
PRO users receive $2 worth of Inference credits monthly that can be used across all supported providers.