feat: add MiniMax provider instrumentation

[octo-patch]

Summary

• Adds MiniMax chat model provider instrumentation using the OpenAI-compatible API (api.minimax.io)
• Detects MiniMax clients by inspecting the base_url of an OpenAI client before falling through to the standard OpenAI path
• Introduces ProviderType.MINIMAX and tags all spans as MINIMAX_API_CALL with model names prefixed as minimax/<model>
• Supports sync/async clients and both streaming and non-streaming calls
• Adds 14 unit tests covering: client detection, span creation, span kind, model name prefix, token usage, error handling, and return value passthrough

Usage

from openai import OpenAI
from judgeval import Tracer, wrap

Tracer.init(project_name="my-project")

client = wrap(OpenAI(
    api_key="MINIMAX_API_KEY",
    base_url="https://api.minimax.io/v1",
))

response = client.chat.completions.create(
    model="MiniMax-M2.7",
    messages=[{"role": "user", "content": "Hello!"}],
    temperature=1.0,
)

Spans will appear in your Judgment dashboard tagged as MINIMAX_API_CALL with model minimax/MiniMax-M2.7.

Supported models

Model ID	Description
MiniMax-M2.7	Peak Performance. Ultimate Value.
MiniMax-M2.7-highspeed	Same performance, faster and more agile

API reference

• Chat (OpenAI Compatible): https://platform.minimax.io/docs/api-reference/text-openai-api

---

[propel-code-bot]

Review found no issues; MiniMax instrumentation appears well-covered and ready to merge.

Status: No Issues Found | Risk: Low

Review Details

📁 10 files reviewed | 💬 0 comments

[propel-code-bot]

✔️ Propel has finished reviewing this change.