whichdevtool?
Menu
StacksDirectory Pricing changes Glossary
VPS and cloud servers App hosting and PaaS Managed databases Transactional email APIs Email marketing and newsletters Domain registrars CDN, DNS and edge security Monitoring, uptime and error tracking Auth and identity Frontend and full-stack frameworks Backend frameworks UI component libraries AI model APIs File and object storage Payments and billing Background jobs, queues and workflows ORMs and database toolkits

Inference provider (EU-hosted open-weight models) · by Scaleway

Scaleway Generative APIs

French cloud's OpenAI-compatible API for open-weight models hosted in Paris, priced in euros, with 1M free tokens for new customers.

Visit scaleway.com ↗ Alternatives Trial EU region
Scaleway Generative APIs homepage
§1

Overview

Scaleway is a French cloud provider, and Generative APIs is its serverless inference service for open-weight models. Models run in its Paris region behind an OpenAI-compatible API, and the line-up covers Mistral, Qwen, DeepSeek, Gemma and GLM text models, Whisper transcription and embeddings. Scaleway states it does not collect, read, reuse or analyse the content of prompts or outputs.

Prices are in euros per million tokens: Mistral Small 3.2 at €0.15 in and €0.35 out, DeepSeek V4 Flash at €0.40 / €0.80, Mistral Medium 3.5 at €1.50 / €7.50. The Batches API costs half and is not rate limited. New customers get 1,000,000 free tokens and 60 minutes of transcription, consumed from the most expensive tokens first.

There is no hard spending cap, only billing alerts, though rate limits bound how fast you can spend. Official rate limits apply after you add a payment method and rise when you verify your identity; dedicated deployments remove shared limits if you need fixed capacity and fixed billing.

§2

Pricing and plans

Plan Price What you get
mistral-small-3.2-24b-instruct-2506 €0.15 in / €0.35 out per 1M Serverless
gemma-4-26b-a4b-it €0.25 in / €0.50 out per 1M Serverless
deepseek-v4-flash-0731 €0.40 in / €0.80 out per 1M Cached input €0.08
qwen3.8-27b €0.60 in / €3.30 out per 1M Cached input €0.12
mistral-medium-3.5-128b €1.50 in / €7.50 out per 1M Serverless
glm-5.2 €1.80 in / €5.50 out per 1M Serverless
whisper-large-v3 €0.003 per minute Audio transcription
Batches API 50% off No rate limit on batch requests

Prices in EUR per 1M tokens, read 2026-09-29. The free tier is 1,000,000 tokens and 60 minutes of audio transcription for each new customer (applied to the most expensive tokens first), not a monthly allowance, so it is recorded here rather than as a permanent free tier. Minimum billing unit is 1,000 tokens. You cannot set a hard spending cap; billing alerts are available, and total spend is bounded by rate limits. Standard rate limits need a payment method, and they rise automatically once you verify your identity. Embedding models cost €0.10 per 1M tokens.

Source: www.scaleway.com/en/pricing/model-as-a-service/ · prices as published, excluding tax.

§3

What it does

EU hosting
Models served from Scaleway's Paris region by a French provider.
No reuse of content
Scaleway says it does not collect, read, reuse or analyse prompts or outputs.
Batches API
50% discount and no rate limit for non-real-time workloads.
OpenAI-compatible
Works with OpenAI client libraries and LangChain.
Dedicated deployments
Same APIs on provisioned capacity for fixed billing and no shared rate limits.
§4

Strengths and limitations

Strengths

  • EU-hosted inference from an EU company.
  • Clear statement that prompts and outputs are not reused.
  • Euro pricing avoids currency swings for European customers.
  • Half-price Batches API without rate limits.

Limitations

  • The 1M free tokens are a one-off allowance for new customers, not monthly.
  • No hard spending cap; only billing alerts.
  • Smaller model catalogue than large US inference providers.
  • Identity verification is needed for the higher standard rate limits.
§5

Who it suits

Good for

  • European companies that need EU data processing
  • RAG apps using EU-hosted embeddings and open models
  • Batch processing jobs at half price

Look elsewhere if

  • Teams needing proprietary frontier models
  • Anyone who needs a hard spend cap
§6

Alternatives to Scaleway Generative APIs

Tool From Free option Stages
Gemini API AI model API (model maker)
Usage-based Free tier
DeepInfra Inference provider (open-weight models)
Usage-based Paid
Replicate Model hosting API (image, video, audio and LLMs)
$0.09/hr Trial
Mistral API (Mistral Studio) AI model API (model maker)
Usage-based Free tier
GroqCloud Inference provider (open-weight models)
Usage-based Free tier
OpenRouter Model router (unified API)
— Free tier

All AI model APIs →

§7

Questions

Does Scaleway Generative APIs have a free tier?

New customers can process 1,000,000 tokens and 60 minutes of audio transcription for free; after that usage is billed per million tokens.

Where are Scaleway's AI models hosted?

In Scaleway's Paris region, according to its Generative APIs pricing.

Can I set a spending limit on Scaleway Generative APIs?

Not a hard cap. Scaleway offers billing alerts, and dedicated deployments give fixed billing.

Sources read for this page