MiniMax-M3 API Pricing & Cost Calculator

How this is calculated

MiniMax-M3 is priced at $0.30 per million input tokens and $1.20 per million output tokens, with an 80% prompt caching discount ($0.06 per million cached reads). A 50% batch discount is also published on OpenRouter ($0.15 per million inputs, $0.60 per million outputs) - one of the few labs to expose batch directly via the OR API.

Verdict

MiniMax-M3 sits at $0.30/M input and $1.20/M output, with the 80% cache discount dropping cached input to $0.06/M and a published 50% batch discount (on OpenRouter) dropping both rows to $0.15/M input and $0.60/M output. The math against Gemini 3.6 Flash is roughly 3x cheaper on output, and the 1M context window matches MiniMax-M3's larger-tier rivals. The cache hit rate decides most of the per-call cost: a high-volume utility pipeline with stable prompts (classification, routing, formatting) sees input collapse to $0.06/M, and where OpenRouter's batch discount applies, the per-call total falls by another half. This is one of the few labs that exposes batch directly via the OR API, which is what unlocks the half-price bulk path.