Verified 2026-11-18 · rates follow the live catalog

MiniMax M3 API

MiniMax M3 is available through the Openference catalog on the same endpoint as every other model.

Specifications and pricing

MiniMax M3 — published rates in USD
ProviderMinimax
Context window1.0M
Input / 1M tokens$0.30
Output / 1M tokens$1.20
Cached input / 1M tokens$0.06
Off-peak90% off at every hour
Serving precisionFP8
Reasoning tiersadaptive

What a request costs

A call with 50,000 input tokens and 10,000 output tokens bills at $0.027: 50,000 ÷ 1,000,000 × $0.30 for input, plus 10,000 ÷ 1,000,000 × $1.20 for output. A prompt prefix that hits the cache bills at $0.06 per million instead of the full input rate, so a long reused system prompt is far cheaper on the second call than the first. The 90% off at every hour discount applies to requests outside the peak window, so batch work scheduled into those hours costs measurably less.

Calling it

Endpoint

Both API formats are served from one host and one key, so a tool that speaks either works without a translation layer:

Set OPENAI_BASE_URL or ANTHROPIC_BASE_URL to that host and keep the client library you already use.

Send MiniMax M3 in the model field.

Capabilities: image input, tool and function calling, reasoning with adaptive effort tiers.

Other MiniMax models

Every MiniMax model is compared on the MiniMax API pricing page.

The rest of the catalog, with rates side by side, is on all models.

Model catalog · Pricing · Comparisons · Cost guides · Integration guides · FAQ
Guides: GLM coding plan alternative · Free AI API · LLM gateway · LLM proxy · Cheap AI API · OpenAI-compatible API
Model APIs: GLM API · DeepSeek API · Kimi API · Qwen API · MiniMax API