Wingless_Imp_8B: benchmark results

SicariusSicariiStuff's Llama 3/3.1 8B merge (Impish_Mind, Unaligned_BETA, Dusk_Rainbow) tuned for roleplay and creative writing. Provider: Other. Released 2025-01-24. Access: Open.

Unified ELO 1498 ± 1, rank #692 of 1392 rated models, from 10 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Open LLM Leaderboard - IFEval74.3Score89.3
UGI - Willingness (W/10)6.8W/10 Score67
Open LLM Leaderboard - MMLU-Pro29.32Score56.3
Open LLM Leaderboard - MuSR10.89Score55.5
Open LLM Leaderboard - MATH Level 512.01Score54
Open LLM Leaderboard - BBH30.59Score53.9
UGI Leaderboard33.47UGI Score45.1
Open LLM Leaderboard - GPQA4.36Score37.6
UGI - Writing23.51Writing Score19.3
UGI - Natural Intelligence11NatInt Score8.1

Interactive version: theaggregate.ai/model?slug=wingless-imp-8b · How It Works · Data refreshed daily, snapshot 2026-09-05.