Llama 3 8B Lima: benchmark results

Provider: Meta. Access: Open.

Unified ELO 1449 ± 20, rank #1742 of 2928 rated models, from 12 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Open LLM Leaderboard v1 - MMLU64.59Accuracy (%) (5-shot)73.1
Open LLM Leaderboard v1 - GSM8K40.64Accuracy (%) (5-shot)56
Open LLM Leaderboard v1 - HellaSwag82.25Normalized accuracy (%) (10-shot)49.8
Open LLM Leaderboard v1 - WinoGrande76.24Accuracy (%) (5-shot)47
Open LLM Leaderboard - IFEval43.71Score46.6
Open LLM Leaderboard v1 - TruthfulQA MC246.81MC2 (%) (0-shot)35.6
Open LLM Leaderboard v1 - ARC Challenge55.8Normalized accuracy (%) (25-shot)34.2
Open LLM Leaderboard - BBH42.96Score29.5
Open LLM Leaderboard - MATH Level 55.06Score28.2
Open LLM Leaderboard - MMLU-Pro26.26Score27.6
Open LLM Leaderboard - MuSR37.13Score24.7
Open LLM Leaderboard - GPQA23.83Score1.1

Interactive version: theaggregate.ai/model?slug=llama-3-8b-lima · How It Works · Data refreshed daily, snapshot 2026-09-23.