Llama 2 Chat 7B — benchmark results

Provider: Meta. Released 2023-07-18. Access: Open.

Unified ELO 922 ± 151, rank #1839 of 1841 rated models, from 7 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
AA Humanity's Last Exam5.81Accuracy (%)42.3
Artificial Analysis Intelligence Index4.26Intelligence Index12.2
AA GPQA Diamond22.73Accuracy (%)1.3
AA MMLU-Pro16.39Accuracy (%)1.2
Epoch AI - Scicode0Score0.2
AA LiveCodeBench0.21Pass@1 (%)0
AA MATH-5005.87Accuracy (%)0

Interactive version: theaggregate.ai/model?slug=llama-2-chat-7b · How It Works · Data refreshed daily, snapshot 2026-07-25.