Muse Spark 1.3 (High): benchmark results

Provider: Meta. Released 2026-09-02. Access: API.

Unified ELO 1750 ± 1, rank #37 of 3078 rated models, from 14 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Korean CSAT 2026 (Easy Mode) - Physics I50Points (out of 50)95.9
Korean CSAT 2026 (Easy Mode) - Society and Culture46Points (out of 50)91.8
Korean CSAT 2026 (Easy Mode) - Life Science I50Points (out of 50)91.2
Korean CSAT 2026 (Easy Mode) - Chemistry I50Points (out of 50)90.8
Korean CSAT 2026 (Easy Mode) - Total442Points (out of 450)89.1
Korean CSAT 2026 (Easy Mode) - Mathematics100Points (out of 100)88.4
Multi-turn Debate (Lechmazur)1600.8Bradley-Terry Rating78.4
Korean CSAT 2026 (Easy Mode) - Korean97Points (out of 100)70.4
Creative Writing (Lechmazur)0.8Mean Score67.3
Korean CSAT 2026 (Easy Mode) - English97Points (out of 100)66
Bug Hunt Bench - VS Code Extension8.7Planted Bugs Fixed (out of 45)46.3
CursorBench 4.033.4Score (%)46.1

Interactive version: theaggregate.ai/model?slug=muse-spark-1-3-high · How It Works · Data refreshed daily, snapshot 2026-09-19.