deepseek-llm-7B-base Benchmark Results

deepseek-llm-7B-base — benchmark results

DeepSeek LLM 7B base checkpoint. Provider: DeepSeek. Released 2023-11-29. Access: Open.

Unified ELO 1284 ± 25, rank #1569 of 1776 rated models, from 16 benchmark results.

Benchmark	Score	Metric	Percentile
Open Chinese LLM - HellaSwag	61.91	Accuracy (%)	65
Open Chinese LLM - WinoGrande	64.01	Accuracy (%)	63.4
Open LLM Leaderboard - GPQA	3.13	Score	28.5
Open Chinese LLM - ARC Challenge	47.1	Accuracy (%)	26.3
InfiBench	25.34	Score (%)	21
Open LLM Leaderboard - BBH	9.77	Score	19.8
Open LLM Leaderboard - MMLU-Pro	8.96	Score	19.6
Open LLM Leaderboard - MuSR	3.76	Score	18.1
Open LLM Leaderboard - IFEval	21.79	Score	14.5
Open LLM Leaderboard - MATH Level 5	1.96	Score	13.8
Open Chinese LLM - GSM8K	11.07	Accuracy (%)	9.5
Open Chinese LLM Leaderboard	45.2	Average Score (%)	8.8

Interactive version: theaggregate.ai/model?slug=deepseek-llm-7b-base · How the rankings work · Data refreshed daily, snapshot 2026-07-22.