LangChain logo

Integration guide · Agent framework (Python)

LangChain on EU-hosted open models

Use ChatOpenAI and OpenAIEmbeddings with Frontière AI's base URL to build LangChain apps and agents on EU-hosted open models. ChatOpenAI(base_url=…, api_key=…) from langchain-openai, or the OPENAI_API_BASE variable.

Tested on 13 October 2026. Chat, streamed usage (stream_usage=True), a tool call and bge-m3 embeddings against the gateway, with langchain-openai and gpt-oss-120b on OVHcloud (bge-m3 has been paused since 20 October 2026; the example now uses qwen3-embedding-8b).

API format
OpenAI-compatible
Base URL
https://getfrontiereai.eu/api/v1
Where it goes
ChatOpenAI(base_url=…, api_key=…) from langchain-openai, or the OPENAI_API_BASE variable.
Your prompts
Straight from your machine to us

Setup

1.Chat model and embeddings

With a custom base URL, LangChain leaves token usage off while streaming: set stream_usage=True. Pin use_responses_api=False so it stays on chat completions. For embeddings, send raw strings instead of tiktoken tokens.

python
# pip install -U langchain-openai
from langchain_openai import ChatOpenAI, OpenAIEmbeddings

llm = ChatOpenAI(
    model="glm-5.2",
    base_url="https://getfrontiereai.eu/api/v1",
    api_key="sk-front-...",
    stream_usage=True,
    use_responses_api=False,
)
print(llm.invoke("Hello from Europe").content)

embeddings = OpenAIEmbeddings(
    model="qwen3-embedding-8b",
    base_url="https://getfrontiereai.eu/api/v1",
    api_key="sk-front-...",
    check_embedding_ctx_length=False,
)

Which model to pick

24 live models have tool calling measured by a real call through the gateway. Models labeled EU sovereign run at European providers; fast-access models are served through US-controlled infrastructure. Prices are per 100,000 tokens, margin and VAT included.

Model idProviderLabelContextInputOutput
pixtral-12bScalewayEU sovereign131k€0.0336€0.0336
mistral-small-3.2-24bScalewayEU sovereign—€0.0252€0.0588
gemma-4-26bAKI.IOEU sovereign256k€0.0168€0.084
gpt-oss-120bScalewayEU sovereign—€0.0252€0.1008
mistral-small-4-119bMistral AIEU sovereign262k€0.0278€0.1109
deepseek-v4-flash-0731ScalewayEU sovereign256k€0.0672€0.1344
qwen3-coder-30bScalewayEU sovereign262k€0.0336€0.1344
llama-3.3-70bScalewayEU sovereign—€0.1512€0.1512
qwen3.6-35bScalewayEU sovereign262k€0.042€0.252
deepseek-v4.1-flashAKI.IOEU sovereign1M€0.1008€0.3024
qwen3-235bScalewayEU sovereign—€0.126€0.378
apertus-v1.5-70bInfomaniakEU sovereign100k€0.1294€0.462
qwen3.8-27bScalewayEU sovereign262k€0.1008€0.5544
qwen3.5-397bScalewayEU sovereign—€0.1008€0.6048
mistral-large-4Mistral AIEU sovereign524k€0.2514€0.7725
glm-5.2ScalewayEU sovereign256k€0.3024€0.924
glm-5.3AKI.IOEU sovereign524k€0.3024€0.924
gemma-3-27bOpenRouterFast access (US)131k€0.0135€0.074
qwen3.8-flashOpenRouterFast access (US)—€0.0252€0.0773
glm-5.3-flashModalFast access (US)1M€0.0689€0.2268
deepseek-v4-proOpenRouterFast access (US)—€0.2453€0.5124
muse-spark-1.1OpenRouterFast access (US)—€0.205€0.6956
muse-spark-1.2OpenRouterFast access (US)—€0.205€0.6956
kimi-k3ModalFast access (US)1M€0.4536€2.268

Good to know

  • ChatOpenAI follows the official OpenAI response format only: reasoning fields such as reasoning_content are dropped.
  • Tool calling and structured output work through ChatOpenAI and depend on the model; see the table below.

Where your prompts go

The library runs in your process and calls the base URL directly. LangSmith tracing is opt-in and would send traces to LangSmith if you enable it.

On our side, a model labeled EU sovereign runs at a European provider, and the gateway itself runs in Paris on Vercel, a US company — our privacy page says so. Billing keeps the model, the token counts and the amount, never the prompt or the reply.

Sovereignty →Privacy →

Sources

Official documentation, checked on 12 October 2026:

FAQ

Does LangChain work with Frontière AI?

Yes. ChatOpenAI(base_url=…, api_key=…) from langchain-openai, or the OPENAI_API_BASE variable. We ran it ourselves: Chat, streamed usage (stream_usage=True), a tool call and bge-m3 embeddings against the gateway, with langchain-openai and gpt-oss-120b on OVHcloud (bge-m3 has been paused since 20 October 2026; the example now uses qwen3-embedding-8b).

Which Frontière AI model should I use with LangChain?

One with measured tool calling, preferably labeled EU sovereign: pixtral-12b, mistral-small-3.2-24b, gemma-4-26b and gpt-oss-120b. The 24 candidates are listed on this page with their context window and price.

Do my prompts stay in Europe with LangChain?

The library runs in your process and calls the base URL directly. LangSmith tracing is opt-in and would send traces to LangSmith if you enable it. On our side, a model labeled EU sovereign runs at a European provider, and the gateway itself runs in Paris on Vercel, a US company — our privacy page says so. Billing keeps the model, the token counts and the amount, never the prompt or the reply.

Other guides

Get a key and try it

Prepaid credit from €10, no subscription. The key works with every guide on this site.

Create an API key

All integrations →

Create an account