AI BriefingScaleway Adds Inference Support for Hugging Face
Key point
Scaleway has been added as a serverless inference provider for Hugging Face, making a variety of open-weight models available via API.
Details
Scaleway has joined Hugging Face Hub as a new Inference Provider. Developers can now use Scaleway's serverless inference service directly from Hugging Face model pages, with easy integration via the Python and JS SDKs.
Key supported models include gpt-oss, Qwen3, DeepSeek R1, and Gemma 3.
Key features of the Scaleway service include:
- Europe-based infrastructure: Uses data centers in Paris, France to ensure data sovereignty and provide low latency for European users.
- High performance: Supports first-token response times of under 200ms, making it suitable for interactive applications and agentic workflows.
- Advanced features: Supports structured outputs, function calling, and text and image multimodal capabilities.
- Reasonable pricing: Offers competitive pricing starting at €0.20 per million tokens.
Users can choose to either use their own Scaleway API key directly, or route billing and requests through Hugging Face.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.