Compare AI models

Specifications
Parameters
1B/4B/12B/27B1.5B/7B/72B
Context window
—4k
API pricing
Cheapest input
$0.04DeepInfra—
Cheapest output
$0.05Parasail—
$10.50Parasail · bf16—
Benchmarks
MATH
89%84%
GSM8K
95.9%96.7%
Overview
CompanyGoogleQwen
Release dateMar 12 2025Aug 8 2024
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

Gemma 3 and Qwen2-Math are evenly matched across the 2 benchmarks they both report (MATH, GSM8K). Qwen2-Math shipped 216 days before Gemma 3, so benchmark comparisons should account for the intervening progress.

Gemma 3 has 1B/4B/12B/27B parameters, while Qwen2-Math has 1.5B/7B/72B.

On MATH, Gemma 3 leads at 89% vs Qwen2-Math at 84%. On GSM8K, Qwen2-Math leads at 96.7% vs Gemma 3 at 95.9%.