MiniMax: MiniMax M2 API Price, Cost Calculator & Specs

At $0.24/1M input and $0.96/1M output tokens, MiniMax-M2's pricing is genuinely disruptive. While most near-frontier models cost $2-5/1M input tokens, this delivers comparable intelligence at a fraction of the cost. The 10B activated parameters (out of 230B total) aren't just an engineering trick - they translate directly to real economics that don't screw up your unit economics.

Here's the raw math: if you're processing 1M input tokens monthly, you're looking at $0.24 instead of the typical $2-5. That's not a discount - that's a paradigm shift for large-scale deployment. The output cost multiplier is worth noting ($0.96 vs typical $1-2), but even with the 4x difference, you're still coming out way ahead compared to alternatives.

What makes this particularly valuable is the focus on coding workflows. The model excels at SWE-Bench Verified, multi-file editing, and compile-run-fix loops - the exact scenarios where you need consistent, reliable execution rather than creative reasoning. The 204K context window handles real development scenarios without breaking a sweat.

Best suited for high-volume developer tools and automated coding workflows where you need reliable execution at scale. The economics make it viable for production deployment with thousands of users, something that becomes impossible with $5+ models. If your use case generates very long responses, pairing it with a cheaper model for summarization makes the math work even better.