How to Get a Free OVHcloud AI Endpoints API Key (2026)

7 free models available — no credit card required. Get your OVHcloud AI Endpoints API key →

Overview

EU-hosted free LLM APIs — GDPR compliant, no registration for anonymous tier.

OVHcloud AI Endpoints provides free API access to open-weight models (Qwen, Mistral, Llama, DeepSeek) hosted in European data centers. The anonymous tier requires no registration at all — just send requests. Registered users get higher rate limits. Ideal for EU developers who need GDPR-compliant hosting and low-latency European inference.

  • No registration needed for anonymous tier
  • EU-hosted (GDPR compliant)
  • Qwen3-Coder, Mistral, Llama, DeepSeek
  • OpenAI-compatible endpoint

API Compatibility: OpenAI SDK-compatible (Chat Completions)

Quick Start Guide

  1. 1
    Go to endpoints.ai.cloud.ovh.net No registration needed for anonymous tier.
  2. 2
    (Optional) Create an OVHcloud account for higher limits
  3. 3
    Pick a model Qwen3-Coder, Mistral, Llama, DeepSeek available.
  4. 4
    Start making requests Anonymous tier: just send requests. Registered: get API token.
  5. 5
    Configure OpenAI client Base URL: https://oai.endpoints.kepler.ai.cloud.ovh.net/v1

All Free OVHcloud AI Endpoints Models — Context Windows & Rate Limits

Model Context Max Output Modality Rate Limit Released Status
Meta-Llama-3_3-70B-Instruct 131K 4K text 2 RPM (anonymous) Online Details
DeepSeek-R1-Distill-Llama-70B 131K 32K text 2 RPM (anonymous) Online Details
Qwen3-Coder-30B-A3B-Instruct 262K 32K textcode 2 RPM (anonymous) Online Details
Qwen2.5-VL-72B-Instruct 128K 8K textimage 2 RPM (anonymous) Online Details
Mistral-Nemo-Instruct-2407 128K 4K text 2 RPM (anonymous) Online Details
Qwen3Guard-Gen-8B 32K 4K text 2 RPM (anonymous) Online Details
Qwen3Guard-Gen-0.6B 32K 4K text 2 RPM (anonymous) Online Details

Pricing & Limits

Credit Card Not required
Free Tier Permanently free
Context Range 32K – 262K
Total Models 7 free
Rate Limits 2 RPM (anonymous)
API Compatibility OpenAI SDK-compatible (Chat Completions)

Use Cases

What OVHcloud AI Endpoints's free models are best for, based on aggregated model capabilities:

Chat 7 models Reasoning 1 model Coding 1 model

Limitations & Caveats

  • Anonymous tier has very low, unpublished rate limits
  • Some models experience cold starts (5-10s first request)
  • EU-hosted only — higher latency outside Europe
See our FAQ for common questions about free LLM APIs