Sociaro

minimax/minimax-m3

MiniMax M3.

Flagship reasoning/coding model, 1M-token context.

A prompt over 512k bills the WHOLE request at the vendor's long-context tier.

Slugminimax/minimax-m3
Kindchat
Vendorminimax
EndpointPOST /v1/chat/completions

Charged per job from what the generator reports it rendered — your Logs show the exact figure for each one; see balance and billing.

Calling it

curl https://api.sociaro.com/v1/chat/completions \
  -H "Authorization: Bearer $SOCIARO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
        "model": "minimax/minimax-m3",
        "messages": [{"role": "user", "content": "hello"}]
      }'

Parameters

FieldTypeRequiredDefaultValuesWhat it does
messagesobject[]yes—OpenAI chat messages (system, user, assistant, tool)The conversation, in OpenAI's format.
max_completion_tokensintegernothe model's1 up to the model's output limitThinking tokens count toward it: too small a cap ends with `finish_reason: length` and an empty answer.
max_tokensintegerno—as aboveThe older name for the same cap; MiniMax marks it deprecated.
streambooleannofalsetrue · falseServer-sent events, as in OpenAI's API.
stream_optionsobjectno—{"include_usage": true}Adds a final chunk with token usage.
temperaturenumberno1(0, 1]MiniMax documents it as having no effect on M3 (fixed at 1).
top_pnumberno0.95(0, 1]MiniMax documents it as having no effect on M3 (fixed at 0.95).
toolsobject[]no—OpenAI function toolsFunction calling, in OpenAI's format.
tool_choicestringnoautonone · auto · requiredWhether the model may, must, or must not call a tool, as MiniMax documents for M3.
thinkingobjectnoon{"type": "adaptive"} · {"type": "disabled"}The model reasons before answering by default; `disabled` turns it off.
reasoning_splitbooleannothe vendor'strue · falsetrue returns the reasoning in `reasoning_content`; false leaves it inside the answer between think tags.
service_tierstringnostandardpriorityMiniMax's priority tier, on M3 only, billed at 1.5× the standard rates.