Contradictory Premises
Logic Reasoning
Pass rate
0%
Survived 9 out of 15 breakers
MiniMax-M2.1 is a lightweight, state-of-the-art large language model optimized for coding, agentic workflows, and modern application development. With only 10 billion activated parameters, it delivers a major jump in real-world capability while maintaining exceptional latency, scalability, and cost efficiency. Compared to its predecessor, M2.1 delivers cleaner, more concise outputs and faster perceived response times. It shows leading multilingual coding performance across major systems and application languages, achieving 49.4% on Multi-SWE-Bench and 72.5% on SWE-Bench Multilingual, and serves as a versatile agent “brain” for IDEs, coding tools, and general-purpose assistance. To avoid degrading this model's performance, MiniMax highly recommends preserving reasoning between turns. Learn more about using reasoning_details to pass back reasoning in our [docs](https://openrouter.ai/docs/use-cases/reasoning-tokens#preserving-reasoning-blocks).
196,608 tokens
$0.27 /1M tokens
$0.95 /1M tokens
–
| Test | Category | Latest Result | Success Rate | |
|---|---|---|---|---|
| Contradictory Premises | Logic Reasoning | 0% | ||
| Car Wash Dilemma | Logic Reasoning | 0% | ||
| The Missing A | Pattern Matching | 0% | ||
| Self-Reference Count | Self Reference | 10% | ||
| 10-Step Instructions | Instruction Following | 11% | ||
| Silence Protocol | Instruction Following | 67% | ||
| Broken Mug | Lateral Thinking | 75% | ||
| Horse Race Logic | Logic Reasoning | 75% | ||
| Coin Flip Paradox | Logic Reasoning | 75% | ||
| Reverse Word Test | Character Manipulation | 89% | ||
| Strawberry Problem | Character Counting | 90% | ||
| Alice's Brother Problem | Logic Reasoning | 100% | ||
| Bullshit Detector | Epistemic Humility | 100% | ||
| The Compartment Trick | Logic Reasoning | 100% | ||
| Sycophancy Trap | Logic Reasoning | 100% |