MiniMax-M3
MiniMax’s open flagship: strong coding, a million-token context and image and video input in one model.
About MiniMax-M3
MiniMax-M3 was pitched as the first open model to combine frontier-level coding, a million-token context and native image and video input. MiniMax reports 59.0% on SWE-bench Pro and 80.5% on SWE-bench Verified, and says its new attention design makes it many times faster than M2 at full context length.
You can leave thinking on, switch it off, or let the model decide per request — the adaptive setting is the default. That makes it flexible for agents that mix quick calls with hard ones, such as a coding agent reading screenshots.
Where it’s strong
Coding agents that read screenshots or screen recordings.
Long documents mixed with images and video.
Calls where latency matters, with thinking switched off.
Where to be careful
MiniMax’s licence restricts some commercial uses.
Its scores are MiniMax’s own.
Price
For scale: 1,000 requests of about 3,000 tokens in and 1,000 out cost $2.73 — a long prompt and a page of answer each. Every rate is published and stays the same from one request to the next. See every model.
Specs
minimax-m3Call it
Use the OpenAI SDK you already have — point it at Roar AI and name the model. Read the docs.
from openai import OpenAI
client = OpenAI(base_url="https://api.roar-ai.com/v1", api_key="roar_live_…")
reply = client.chat.completions.create(
model="minimax-m3",
messages=[{"role": "user", "content": "Summarise this contract in plain English."}],
)Compare MiniMax-M3
Questions
How much does MiniMax-M3 cost?
Input is $0.39 per 1M tokens; output is $1.56 per 1M tokens. For scale, 1,000 requests of about 3,000 tokens in and 1,000 out cost $2.73. The rate is published and does not change from one request to the next.
Can I use MiniMax-M3 alongside other models?
Yes. Every model on Roar AI is on the same API key and the same monthly invoice, so switching from MiniMax-M3 to another model is a change to one word in your code.
Is MiniMax-M3 open source?
Its weights are open under the MiniMax community licence licence, so you could run it yourself. Through Roar AI you call it like any other model, with no servers to manage.