MiniMax M3 API：OpenAI 兼容接入与长文费用

curl --fail-with-body --max-time 120 https://api.apizt.com/v1/chat/completions \
  -H "Authorization: Bearer $tokenba_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"MiniMax-M3","max_tokens":4096,"messages":[{"role":"user","content":"Review a retry policy that retries every failure three times. Explain which HTTP status codes should not be retried and propose exponential backoff with jitter. Do not assume a service SLA."}]}'

import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["tokenba_API_KEY"],
    base_url="https://api.apizt.com/v1",
    timeout=120.0,
)
result = client.chat.completions.create(
    model="MiniMax-M3",
    messages=[{"role": "user", "content": "Turn this project discussion into decisions and open questions: We will add pagination to the model list. The API page size is undecided. Existing deep links must continue working. Do not invent owners or dates."}],
    max_tokens=4096,
)
choice = result.choices[0]
print(choice.message.content)
print("finish_reason:", choice.finish_reason)
print("usage:", result.usage)
if not choice.message.content or choice.finish_reason == "length":
    raise RuntimeError("Check the output budget and channel settings before retrying")