Qwen2.5-14B-Instruct-1M Benchmark Results

Qwen2.5-14B-Instruct-1M — benchmark results

Alibaba's long-context variant of Qwen2.5 14B Instruct, extending input length to 1M tokens. Provider: Alibaba. Released 2025-01-27. Access: Open.

Unified ELO 1549 ± 23, rank #621 of 1776 rated models, from 78 benchmark results.

Benchmark	Score	Metric	Percentile
Open LLM Leaderboard - IFEval	84.14	Score	99.6
Open LLM Leaderboard - MATH Level 5	53.02	Score	98.2
EVALITA - text-entailment	85.02	CPS	97.9
RULER	95.7	Avg Accuracy (%)	95
Open Japanese LLM - CG	63.86	Score (%)	94.8
Open Japanese LLM - Jsick Exact Match	84.84	Score (%)	90.4
EVALITA - relation-extraction	44.23	CPS	89.4
Open Japanese LLM - Jsts Spearman	87.81	Score (%)	88.7
Open Japanese LLM - Wiki Dependency SET F1	41.23	Score (%)	88.2
Open Japanese LLM - Mbpp Pylint Check	95.78	Score (%)	87
Open LLM Leaderboard - MMLU-Pro	42.77	Score	86.8
Open LLM Leaderboard - BBH	45.66	Score	85.8