OVHcloud AI Endpoints
by OVH Groupe SA · www.ovhcloud.com ↗
Serverless inference API from OVHcloud that serves open-weight language, vision, embedding, image and speech models through OpenAI-compatible endpoints, billed by usage.
cloud inference serverless open-models europe
- Category
- Platforms & Infrastructure
- Business model
- Freemium
- Availability
- Global
- Launched
- 2025
- Record updated
- 2026-10-12
- Pricing
- www.ovhcloud.com ↗
- Documentation
- docs.ovhcloud.com ↗
- Founded
- 1999
- Headquarters
- FR
Selection profile
Every field says where its value comes from. "Unknown" means we have not confirmed it — it does not mean the product lacks the feature.
- Tasks it fits
- Running models yourself
Derived from the record: category "platforms".
Suggest a correction - How you use it
- API
Matched on the vendor product page: "AI Endpoints is the serverless inference API for a selection of models".
Suggest a correction - Runs on
- Web
Derived from platform words in the record description/features/tags.
Suggest a correction - Where it runs
- Vendor cloud
The vendor states the service runs "on OVHcloud's robust and confidential infrastructure"; it is reached over the network, not installed on the user's machine.
Suggest a correction - Works offline
- Not recorded
The record contains no offline-operation claim.
Suggest a correction - Official service region
- Global
Mirrors the availability field already on the record ("global"); the vendor says it serves 140 countries, and no per-region service list was read.
Suggest a correction - Public API
- Yes
The vendor documentation documents OpenAI-compatible HTTP routes, including /v1/chat/completions and /v1/responses.
Suggest a correction - Licence
- proprietary
Commercial product; no source licence is published in the record.
Suggest a correction - Commercial use
- Not recorded
Commercial product; the pages read do not state a non-commercial restriction, but that was not verified.
Suggest a correction - Interface languages
- Not recorded
No vendor page read in this pass states the product's own interface languages; the key exists so the field can be answered later rather than left absent.
Suggest a correction - Billing
- Usage-based Free allowance: Limited Subscription required: No
Prices are the vendor’s advertised entry price and exclude taxes; always confirm on the vendor’s own page.
Suggest a correction
Where each value comes from
- Tasks it fits
- Our inference
- How you use it
- Verified 2026-10-12 www.ovhcloud.com ↗
- Runs on
- Our inference
- Where it runs
- Verified 2026-10-12 www.ovhcloud.com ↗
- Works offline
- Our inference
- Official service region
- Our inference
- Public API
- Verified 2026-10-12 docs.ovhcloud.com ↗
- Licence
- Our inference
- Commercial use
- Our inference
- Interface languages
- Unknown
Key facts and their evidence
Each fact lists the conditions it holds under, the page it was read from and the day it was checked.
-
Headquarters: FR Verified 2026-10-12 www.ovhcloud.com ↗
Applies to: As stated on ovhcloud.com/en/about-us/ on 2026-10-12
Matched on the vendor about page: "Founded in 1999, OVHcloud is a French company with an international presence". The Public Cloud terms on the AI Endpoints page name "OVH Groupe SA (537 407 926 RCS Lille Métropole, 2 rue Kellermann, 59100 Roubaix, France)".
Suggest a correction -
Pricing: Pay-as-you-go per million tokens, per-model rates published; US$200 Public Cloud free-trial credit Verified 2026-10-12 www.ovhcloud.com ↗
Applies to: As stated on ovhcloud.com on 2026-10-12
The product page offers "US$ 200 in free credit" and the catalogue lists per-model prices in €/Mtoken (e.g. embeddings 0.01 €, a large visual model 0.91 €), with a few models shown as "Free". No periodic plan price is published.
Suggest a correction -
API: OpenAI-compatible /v1/chat/completions and /v1/responses routes Verified 2026-10-12 docs.ovhcloud.com ↗
Applies to: As documented at docs.ovhcloud.com on 2026-10-12
The documentation describes /v1/responses as "the most recent OpenAI-compatible route" and, like v1/chat/completions, usable for text generation, multi-turn conversations and tool calling.
Suggest a correction
Compare this tool Suggest a correction
Overview
OVHcloud AI Endpoints is a serverless inference service operated by OVH Groupe SA, the French cloud provider founded in 1999 and headquartered in Roubaix. It gives developers OpenAI-compatible API access to more than 40 open-weight models — Llama, Qwen, Mistral, gpt-oss, Whisper and Stable Diffusion among them — covering chat, reasoning, vision, embeddings, image generation, transcription and moderation without requiring users to deploy GPUs. OVHcloud launched the service on 29 April 2025 and bills it pay-as-you-go per million tokens, with a US$200 Public Cloud trial credit for new projects. The limitation is on the newest route: OVHcloud's own documentation says the OpenAI-compatible /v1/responses endpoint has no statefulness and no built-in tools such as web search, so callers must send store: false and rely on custom function tools.
Key features
- Serverless, OpenAI-compatible inference: the /v1/chat/completions and /v1/responses routes
- More than 40 models across LLMs, visual LLMs, embeddings, image generation, speech-to-text and text-to-speech
- Virtual models that map a request to the best available model by size, price or capability
- Zero data retention: OVHcloud states customer data is never used to train or improve its models
- Per-model pricing published in the catalogue, in euros per million tokens
- Model lifecycle management with transparent versioning, hosted on OVHcloud's own European infrastructure
Use cases
- Adding chat, reasoning or coding assistance to an application without managing GPUs
- Getting embeddings, images or speech from the same OpenAI-compatible key
- Keeping inference inside European infrastructure for data-residency requirements
Pricing
Billed by usage rather than by subscription: each model's price per million tokens is published in the public catalogue, ranging from around €0.01 for an embedding model to about €0.91 for a large visual model, with a few models (moderation and image generation) listed as free. OVHcloud offers a US$200 Public Cloud free-trial credit to first-time projects. Rates change, so the catalogue is the source of record; this record sets no comparable paid_from figure.
Frequently asked questions
What is OVHcloud AI Endpoints?
A serverless inference API: it exposes pre-trained open-weight models for chat, reasoning, vision, embeddings, image generation and speech over OpenAI-compatible HTTP routes, so an application does not need to run its own GPUs.
Where is OVHcloud based?
In Roubaix, France. OVHcloud's about page calls the company "a French company" founded in 1999, and the Public Cloud terms name OVH Groupe SA at 2 rue Kellermann, 59100 Roubaix.
How is it billed?
Pay-as-you-go per million tokens, with each model's rate published in the catalogue, plus a US$200 free-trial credit for first-time Public Cloud projects. No subscription is required.
What are the limits of the API?
On the /v1/responses route, OVHcloud's documentation says statefulness is not managed (always send store: false) and built-in tools such as web search, file search or code execution are unsupported; only custom function tools work.
What does this record not establish?
No page read for this record states the product's own interface languages, its licence terms, or a per-region service list, so those fields stay unconfirmed.
Similar products
- Gcore — Provider of edge AI, cloud, network, and security services, including GPU cloud, serverless inference with OpenAI-compatible APIs, and a global edge network.
- Infomaniak AI Services — Swiss-hosted AI service that serves open-source language, transcription and image models through an OpenAI-compatible API and a web app, billed on credits.
- Nscale — AI cloud provider offering GPU clusters, managed Kubernetes and Slurm, and a serverless inference API compatible with OpenAI SDKs.
- Scaleway — French cloud provider whose platform includes NVIDIA GPU instances, managed Kubernetes through Kapsule, and EU-hosted Generative APIs compatible with the OpenAI SDK.
- STACKIT AI Model Serving — Managed STACKIT service that serves open-source large language models through an API inside German data centres, with token-based billing and no storage of queries.
- Nebius — AI cloud platform for workloads from training through inference, with NVIDIA GPU instances billed per GPU-hour, managed Kubernetes and Slurm, and token-based inference services.
All OVHcloud AI Endpoints alternatives → · All Platforms & Infrastructure products →
Related guides
Sources
- OVHcloud AI Endpoints ↗
- About us ↗
- AI Endpoints model catalogue and prices ↗
- Public Cloud price list ↗
- AI Endpoints — Responses API (documentation) ↗
- OVHcloud launches AI Endpoints (press release, 29 April 2025) ↗
- OVHcloud — Wikipedia ↗
This record was last reviewed on 2026-10-12.
Machine-readable record: /api/products/ovhcloud-ai-endpoints.json · Spot an error? Suggest a correction