Moonshot AI·released 2025-07-11
Scored on 4 benchmarks, leading 1. 0% of its scores are independently reproduced — the rest are self-reported or unverified.
Leads on: Lech Mazur Writing
Every score, ranked, colored by evidence status.
| Benchmark | Score | Evidence status | Measured |
|---|---|---|---|
| Lech Mazur Writing | 87.29accuracy (%) | unverified | 2025-07-11 |
| Fiction.LiveBench | 66.7accuracy (%) | unverified | 2025-07-11 |
| Aider polyglot | 59.1accuracy (%) | unverified· optimized | 2025-07-11 |
| WeirdML | 36.68accuracy (%) | unverified | 2025-07-11 |