Implementation:Mistralai Client python GCP Chat And Fim Complete

Knowledge Sources	Mistral Client Python Google Cloud Vertex AI
Domains	Cloud_Deployment, GCP, LLM_Inference, Code_Generation
Last Updated	2026-02-15 14:00 GMT

Overview

Concrete tool for chat completion and fill-in-the-middle code completion on GCP Vertex AI provided by the mistralai_gcp Chat and Fim resources.

Description

The GCP Chat resource provides complete(), complete_async(), stream(), and stream_async() — identical to the standard SDK. The Fim resource provides complete() and complete_async() for fill-in-the-middle code generation. All requests are transparently routed to Vertex AI endpoints via URL rewriting. The Fim API takes a prompt (code before cursor) and optional suffix (code after cursor).

Usage

Access chat via client.chat.complete() and FIM via client.fim.complete() on a MistralGoogleCloud client instance.

Code Reference

Source Location

Repository: client-python
File: packages/mistralai_gcp/src/mistralai_gcp/chat.py (L1-681), packages/mistralai_gcp/src/mistralai_gcp/fim.py (L1-533)

Signature

# Chat (same as standard SDK)
class Chat:
    def complete(self, *, model: str, messages: List[...], ...) -> ChatCompletionResponse: ...
    def stream(self, *, model: str, messages: List[...], ...) -> EventStream[CompletionEvent]: ...

# FIM (GCP-only)
class Fim:
    def complete(
        self,
        *,
        model: str,
        prompt: str,
        suffix: Optional[str] = None,
        temperature: Optional[float] = None,
        top_p: Optional[float] = None,
        max_tokens: Optional[int] = None,
        random_seed: Optional[int] = None,
        stop: Optional[Union[str, List[str]]] = None,
    ) -> FIMCompletionResponse: ...

    async def complete_async(self, *, model: str, prompt: str, ...) -> FIMCompletionResponse: ...

Import

from mistralai_gcp import MistralGoogleCloud
# Access via: client.chat.complete(...) or client.fim.complete(...)

I/O Contract

Inputs (Chat)

Name	Type	Required	Description
model	str	Yes	Mistral model name (URL auto-rewritten by GCP hook)
messages	List[ChatCompletionRequestMessage]	Yes	Conversation messages

Inputs (FIM)

Name	Type	Required	Description
model	str	Yes	Mistral code model name
prompt	str	Yes	Code before the cursor (prefix)
suffix	Optional[str]	No	Code after the cursor
max_tokens	Optional[int]	No	Maximum generated tokens

Outputs

Name	Type	Description
ChatCompletionResponse	ChatCompletionResponse	For chat requests
FIMCompletionResponse	FIMCompletionResponse	For FIM requests; same structure as ChatCompletionResponse

Usage Examples

GCP Chat Completion

import asyncio
from mistralai_gcp import MistralGoogleCloud

client = MistralGoogleCloud(region="europe-west4", project_id="my-project")

async def main():
    response = await client.chat.complete_async(
        model="mistral-large-latest",
        messages=[{"role": "user", "content": "What is Vertex AI?"}],
    )
    print(response.choices[0].message.content)

asyncio.run(main())

FIM Code Completion

from mistralai_gcp import MistralGoogleCloud

client = MistralGoogleCloud(region="europe-west4", project_id="my-project")

response = client.fim.complete(
    model="codestral-latest",
    prompt="def fibonacci(n):\n    if n <= 1:\n        return n\n    ",
    suffix="\n\nprint(fibonacci(10))",
)
print(response.choices[0].message.content)

Related Pages

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment