Compare AI models

Specifications
Parameters
—14B
Context window
32k2k
Benchmarks
HumanEval
74.4%32.3%
MATH
53.2%24.8%
MMLU
83.7%66.3%
GSM8K
88.9%61.3%
Overview
CompanyGoogleQwen
Release dateFeb 8 2024Sep 25 2023
AccessClosedOpen Weight
Model detailsView modelView model

Frequently asked questions

Gemini 1.0 Ultra leads Qwen-14B on 4 of the 4 benchmarks they both report (HumanEval, MATH, MMLU, GSM8K). Qwen-14B shipped 136 days before Gemini 1.0 Ultra, so benchmark comparisons should account for the intervening progress.

Context windows are 32k (Gemini 1.0 Ultra) vs 2k (Qwen-14B). Gemini 1.0 Ultra is closed, while Qwen-14B is open weight.

On HumanEval, Gemini 1.0 Ultra leads at 74.4% vs Qwen-14B at 32.3%. On MATH, Gemini 1.0 Ultra leads at 53.2% vs Qwen-14B at 24.8%. On MMLU, Gemini 1.0 Ultra leads at 83.7% vs Qwen-14B at 66.3%. On GSM8K, Gemini 1.0 Ultra leads at 88.9% vs Qwen-14B at 61.3%.