OVHcloud on Hugging Face Inference Providers 🔥

Source: HuggingFace Date: 2025-11-24 URL: https://huggingface.co/blog/OVHcloud/inference-providers-ovhcloud

Summary

Integration announcement: OVHcloud joins HF Inference Providers as a European cloud alternative. Available models include gpt-oss-120b, Qwen3, DeepSeek R1, and Llama. Pricing: €0.04/million tokens with sub-200ms first-token latency from European data centers. Accessible via HF UI, Python SDK, and JS client; two modes — direct (billed to OVHcloud) or HF-routed (billed via HF account). No benchmark numbers.

Implications

HF as open-source ML hub. OVHcloud as the second European provider (after Scaleway) in the HF Inference Providers program gives European organizations a choice of EU-based inference backends through the same SDK. The pricing gap (€0.04/million vs Scaleway’s €0.20/million for comparable models) will drive routing decisions — the HF aggregation layer now has meaningful price competition among European-compliant options.

Open-weights ecosystem health. gpt-oss-120b availability through OVHcloud means the largest publicly available open-weights model is now accessible through multiple EU-compliant inference backends. For regulated industries with GDPR and data residency requirements, having two competing European providers at meaningfully different price points is a significant practical improvement over the previous single-provider situation.