Llama-3.1-Nemotron-lorablated-70B — benchmark results

mlabonne's LoRA-ablated (uncensored) merge of Llama 3.1 Nemotron 70B. Provider: NVIDIA. Released 2024-10-17. Access: Open.

Unified ELO 1625 ± 20, rank #426 of 1776 rated models, from 10 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Open LLM Leaderboard - GPQA18.79Score97.5
Open LLM Leaderboard - BBH54.18Score96.5
Open LLM Leaderboard - MMLU-Pro48.26Score93.7
Open LLM Leaderboard - MuSR18.35Score90.7
Open LLM Leaderboard - IFEval72.29Score86.2
UGI Leaderboard47.35UGI Score84.6
Open LLM Leaderboard - MATH Level 533.38Score84.1
UGI - Writing42.05Writing Score79.9
UGI - Willingness (W/10)7.5W/10 Score77.6
UGI - Natural Intelligence30.29NatInt Score75

Interactive version: theaggregate.ai/model?slug=llama-3-1-nemotron-lorablated-70b · How the rankings work · Data refreshed daily, snapshot 2026-07-22.