Muse Spark 1.1 (High): benchmark results

Provider: Meta. Released 2026-07-09. Access: API.

Unified ELO 1714 ± 1, rank #74 of 1761 rated models, from 7 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Conceptual Reasoning Index - Argument Evaluation (LMCA)49.91Chance-Corrected Score (0-100)91.9
Multi-turn Debate (Lechmazur)1668Bradley-Terry Rating91.1
Epoch AI - Dtbench94.4Score90.3
Creative Writing (Lechmazur)1.1Mean Score71.3
Surface Evolver Bench Pass Rate37.5Pass Rate (%)60
ComplexConstraints36.9Score (%)59.9
Surface Evolver Bench52.5Mean Score (%)38

Interactive version: theaggregate.ai/model?slug=muse-spark-1-1-high · How It Works · Data refreshed daily, snapshot 2026-09-05.