Compare AI models

Specifications
Parameters
—7B/14B
Context window
32k1M
Benchmarks
HumanEval
74.4%88.4%
MATH
53.2%79.5%
GSM8K
88.9%94.8%
Overview
CompanyGoogleQwen
Release dateFeb 8 2024Jan 27 2025
AccessClosedOpen Weight
Model detailsView modelView model

Frequently asked questions

Qwen2.5-1M leads Gemini 1.0 Ultra on 3 of the 3 benchmarks they both report (HumanEval, MATH, GSM8K). Gemini 1.0 Ultra shipped 354 days before Qwen2.5-1M, so benchmark comparisons should account for the intervening progress.

Context windows are 32k (Gemini 1.0 Ultra) vs 1M (Qwen2.5-1M). Gemini 1.0 Ultra is closed, while Qwen2.5-1M is open weight.

On HumanEval, Qwen2.5-1M leads at 88.4% vs Gemini 1.0 Ultra at 74.4%. On MATH, Qwen2.5-1M leads at 79.5% vs Gemini 1.0 Ultra at 53.2%. On GSM8K, Qwen2.5-1M leads at 94.8% vs Gemini 1.0 Ultra at 88.9%.