European LLM APIs
3 of the 14 LLM API providers we track have a parent company in Europe: Mistral AI, Nebius Token Factory and Scaleway Generative APIs. At 500M input tokens / month, 50M output tokens / month, the lowest European estimate is $990/yr for Mistral AI Ministral 3 8B, against $324/yr for the cheapest US-owned option, DeepInfra gpt-oss-120b. All of them offer EU data residency on a self-serve plan.
“European” means the ultimate parent company is in the EU/EEA, the UK or Switzerland. That describes ownership, not where your data is stored or every law that applies. How we define jurisdiction.
European LLM API providers by estimated bill
| Service | Parent company in | EU residency (self-serve) | Est. yearly |
|---|---|---|---|
| Mistral AI Ministral 3 8BMistral AI's La Plateforme API serving Mistral Medium, Large, Small and Ministral models, most of them open-weight, hosted in the EU. | 🇫🇷 France | Yes | $990/yrMinistral 3 8B · $0.15 in / $0.15 out per MPublished rates |
| Nebius Token Factory DeepSeek-V4-Flash-0731Nebius inference API (formerly AI Studio) for open-weight models, with endpoints in EU, UK, US and other regions. | 🇳🇱 NetherlandsNebius Group | Yes | $1,008/yrDeepSeek-V4-Flash-0731 · $0.14 in / $0.28 out per MPublished rates |
| Mistral AI Mistral Small 4Mistral AI's La Plateforme API serving Mistral Medium, Large, Small and Ministral models, most of them open-weight, hosted in the EU. | 🇫🇷 France | Yes | $1,260/yrMistral Small 4 · $0.15 in / $0.6 out per MPublished rates |
| Nebius Token Factory gpt-oss-120bNebius inference API (formerly AI Studio) for open-weight models, with endpoints in EU, UK, US and other regions. | 🇳🇱 NetherlandsNebius Group | Yes | $1,260/yrgpt-oss-120b · $0.15 in / $0.6 out per MPublished rates |
| Scaleway Generative APIs Mistral Small 3.2 24BServerless OpenAI-compatible API for open-weight models, hosted in Paris, France by Scaleway (Iliad group). | 🇫🇷 FranceIliad | Yes | $1,326/yrMistral Small 3.2 24B · $0.18 in / $0.41 out per MPublished rates |
| Scaleway Generative APIs gpt-oss-120bServerless OpenAI-compatible API for open-weight models, hosted in Paris, France by Scaleway (Iliad group). | 🇫🇷 FranceIliad | Yes | $1,500/yrgpt-oss-120b · $0.18 in / $0.7 out per MPublished rates |
| Scaleway Generative APIs DeepSeek-V4-Flash-0731Serverless OpenAI-compatible API for open-weight models, hosted in Paris, France by Scaleway (Iliad group). | 🇫🇷 FranceIliad | Yes | $3,384/yrDeepSeek-V4-Flash-0731 · $0.47 in / $0.94 out per MPublished rates |
| Mistral AI Mistral Large 3Mistral AI's La Plateforme API serving Mistral Medium, Large, Small and Ministral models, most of them open-weight, hosted in the EU. | 🇫🇷 France | Yes | $3,900/yrMistral Large 3 · $0.5 in / $1.5 out per MPublished rates |
| Scaleway Generative APIs Qwen3-235B-A22B-Instruct-2507Serverless OpenAI-compatible API for open-weight models, hosted in Paris, France by Scaleway (Iliad group). | 🇫🇷 FranceIliad | Yes | $6,858/yrQwen3-235B-A22B-Instruct-2507 · $0.88 in / $2.63 out per MPublished rates |
| Nebius Token Factory GLM-5.3Nebius inference API (formerly AI Studio) for open-weight models, with endpoints in EU, UK, US and other regions. | 🇳🇱 NetherlandsNebius Group | Yes | $11,040/yrGLM-5.3 · $1.4 in / $4.4 out per MPublished rates |
| Nebius Token Factory DeepSeek-V4-ProNebius inference API (formerly AI Studio) for open-weight models, with endpoints in EU, UK, US and other regions. | 🇳🇱 NetherlandsNebius Group | Yes | $12,600/yrDeepSeek-V4-Pro · $1.75 in / $3.5 out per MPublished rates |
| Mistral AI Mistral Medium 3.5Mistral AI's La Plateforme API serving Mistral Medium, Large, Small and Ministral models, most of them open-weight, hosted in the EU. | 🇫🇷 France | Yes | $13,500/yrMistral Medium 3.5 · $1.5 in / $7.5 out per MPublished rates |
| Scaleway Generative APIs GLM-5.2Serverless OpenAI-compatible API for open-weight models, hosted in Paris, France by Scaleway (Iliad group). | 🇫🇷 FranceIliad | Yes | $16,524/yrGLM-5.2 · $2.11 in / $6.44 out per MPublished rates |
| Nebius Token Factory Kimi-K3Nebius inference API (formerly AI Studio) for open-weight models, with endpoints in EU, UK, US and other regions. | 🇳🇱 NetherlandsNebius Group | Yes | $27,000/yrKimi-K3 · $3 in / $15 out per MPublished rates |
At 500M input tokens / month, 50M output tokens / month. Price per token isn’t price per task. Cheaper models can need more retries or longer outputs. Benchmark on your own workload. Sources last checked 2026-10-06. Change the usage to see bills for your workload.
US-owned services with EU data residency
These offer EU data residency on a self-serve plan, but their parent company is in the US.
- OpenAI — OpenAI's API serving the GPT-6 family (Astra, Sol, Luna) and older GPT-5.x models, with batch, caching and EU regional processing options.
- Together AI — Serverless and dedicated inference for open-weight models (DeepSeek, Kimi, GLM, Qwen, Llama, gpt-oss) via an OpenAI-compatible API.