Leaderboards

MetaMuse Spark 1.3

meta/muse-spark-1.3
Overall score
3.191
95% CI [3.162, 3.221]
Based on 4,531 sessions
Rank 9 of 13 models on the Chinese leaderboard
4 aspects
● Conversational ability 0%4.11
● Character consistency 100%3.19
● Role-playing attractiveness 0%3.50
● Personality back-testing 0%—
13 dimensions
Fluency
3.91
Coherency
4.22
Consistency
4.19
Knowledge exposure
2.19
Knowledge accuracy
3.13
Knowledge hallucination
3.25
Persona behavior
3.92
Persona utterance
3.48
Human-likeness
4.09
Communication skills
3.16
Expression diversity
3.51
Empathy
3.23
MBTI accuracy
—
FluencyCoherencyConsistencyKnowledge exposureKnowledge accuracyKnowledge hallucinationPersona behaviorPersona utteranceHuman-likenessCommunication skillsExpression diversityEmpathyMBTI accuracy2

Scores by language

Tested November 1, 2026
LanguageRankOverall95% confidence intervalConversational abilityCharacter consistencyRole-playing attractivenessPersonality back-testingSessions
Chinese9 / 133.193.2 – 3.24.13.23.5—4,531