Verdict
Built from the measurements below
- Accuracy is too close to call: MiniMax M3 scored 57.5% and MiniMax M3.1 Flash 66.2% on the same 80 questions, a gap within what chance alone could produce.
- MiniMax M3.1 Flash is 3.0× cheaper per question as billed ($0.00087 vs $0.0026), reasoning tokens included.
- Output speed is similar (121 vs 131 tokens per second).
Which should you choose?
Choose MiniMax M3.1 Flash for the lower cost per correct answer ($0.0013 against $0.0045) and shorter outputs (about 1,376 tokens per question against 2,129). The widest subject gap is in economics: MiniMax M3.1 Flash answered 8 of 9 correctly, MiniMax M3 6 of 9. With so few questions per subject, treat this as a lead to test on your own prompts, not a result.
Accuracy gap -8.8 points (z = -1.14). We call a winner only when z is at least 1.96, the usual 95% threshold. Speed, latency and price need a 10% margin.