M3 and Highspeed side by side
Flagship M3 for agents and multimodal work; M2.7 Highspeed when the loop is latency-bound.
OpenAI-compatible · MiniMax M3 · Agents & multimodalMiniMax APIM3 on your existing client
MiniMax M3 and M2.7 Highspeed through one OpenAI-compatible Base URL. Agent, reasoning, and latency-sensitive SKUs in the same catalog.
Why SupaNexus
MiniMax without a parallel integration.
Same OpenAI client
Chat Completions, streaming, and your existing retries. Only the model ID changes.
Shared governance
Quotas, metering, and zero retention apply to MiniMax the same way they apply to DeepSeek.
Catalog
MiniMax models on SupaNexus
Send the full Public API ID. Highspeed is a distinct ID, not a query parameter.
| Model | Public API ID | Context | Best for |
|---|---|---|---|
| MiniMax M3 | minimax/minimax-m3 | 1M | MultimodalAgentsReasoning |
| MiniMax M2.7 | minimax/minimax-m2.7 | 204.8K | ChatThinking |
| MiniMax M2.7 Highspeed | minimax/minimax-m2.7-highspeed | 204.8K | Low latencyThinking |
Live rates are in the developer console. See also homepage pricing · console catalog
Quickstart
First MiniMax call in 30 seconds
Official OpenAI SDK. Model ID minimax/minimax-m3.
https://api.supanexus.ai/v1
Python
from openai import OpenAI
client = OpenAI(
base_url="https://api.supanexus.ai/v1",
api_key="sk-snx-...",
)
response = client.chat.completions.create(
model="minimax/minimax-m3",
messages=[{"role": "user", "content": "Hello!"}],
)
print(response.choices[0].message.content)
FAQ
MiniMax API questions
M3 vs Highspeed, IDs, and compatibility.
minimax/minimax-m3 is the current flagship (agents, multimodal). minimax/minimax-m2.7-highspeed when you need the faster M2.7 lane.
The full Public API ID, for example minimax/minimax-m3. Do not strip the vendor prefix.
Yes. Point it at https://api.supanexus.ai/v1 with a SupaNexus project key.
Turn on MiniMax M3 with one key.
Create a project, copy the key, send minimax/minimax-m3.