deepseek-moe-16B-chat Benchmark Results

deepseek-moe-16B-chat: benchmark results

Provider: DeepSeek. Released 2024-01-09. Access: Open.

Unified ELO 1328 ± 40, rank #2360 of 2656 rated models, from 12 benchmark results.

Benchmark	Score	Metric	Percentile
Open Portuguese LLM - FaQuAD NLI	61.72	Macro F1 (%)	46.1
Open LLM Leaderboard - IFEval	36.63	Score	35.2
Open Portuguese LLM - ENEM	48.22	Accuracy (%)	32.9
Open Portuguese LLM - OAB Exams	35.26	Accuracy (%)	32.4
Open Portuguese LLM - BLUEX	37.97	Accuracy (%)	31.3
Open LLM Leaderboard - MuSR	38.08	Score	30
Open Portuguese LLM - ASSIN2 RTE	77.68	Macro F1 (%)	29.2
Open LLM Leaderboard - MMLU-Pro	19.64	Score	21.3
Open LLM Leaderboard - MATH Level 5	2.57	Score	16.7
Open LLM Leaderboard - BBH	32.75	Score	15
YALL Nous Leaderboard	44.72	Average	9.5
Open LLM Leaderboard - GPQA	22.48	Score	0.1

Interactive version: theaggregate.ai/model?slug=deepseek-moe-16b-chat · How It Works · Data refreshed daily, snapshot 2026-09-19.