20-30b Dense大模型有哪些好的

Administrator

2026-02-22

19

Qwen3.5-27b 我宣布封神!!!

int6恰好22.5G显存占用

qwen3

幻觉重,工业垃圾

qwen2.5

14b

14b-1M

32b

上下文:128k

qwen2.5-32B-instruct_wturbo.001.jpg

QwQ

幻觉太重,不要

InternLM2.5

上下文:200k

enchmark	InternLM2.5-20B	InternLM2-20B
MMLU	74.25	67.58
CMMLU	82.22	68.29
BBH	77.82	71.36
MATH	48	32.66
HUMANEVAL	71.95	51.22
GPQA	37.88	31.31

GPT-OSS 20b

moe,不要

Gemma3

幻觉太重,不要

GLM-Z1-32B

面向推理,无人问津,不踩这个地雷

Mistral small

视觉多模态

上下文:128k

Model	MMLU	MMLU Pro (5-shot CoT)	MATH	GPQA Main (5-shot CoT)	GPQA Diamond (5-shot CoT )	MBPP Plus - Pass@5	HumanEval Plus - Pass@5	SimpleQA (TotalAcc)
Small 3.1 24B Instruct	80.62%	66.76%	69.30%	44.42%	45.96%	74.63%	88.99%	10.43%
Small 3.2 24B Instruct	80.50%	69.06%	69.42%	44.22%	46.13%	78.33%	92.90%	12.10%