Compare AI models

Specifications
Parameters
14B1.5B/7B/72B
Context window
256k4k
API pricing
Input price
$0.20—
Output price
$0.20—
Cached input price
$0.02—
Cheapest input
$0.20Mistral—
Cheapest output
$0.20Mistral—
$50.00Mistral list price—
Benchmarks
MATH
90.4%85.9%
Overview
CompanyMistralQwen
Release dateDec 2 2025Sep 19 2024
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

Ministral 3 14B-25-12 leads Qwen2.5-Math on 1 of the 1 benchmark they both report (MATH). Only Ministral 3 14B-25-12 has a verified first-party price for an available API: $0.20 per million input tokens and $0.20 per million output tokens. No currently available pay-as-you-go API rate is tracked for Qwen2.5-Math. Qwen2.5-Math shipped 439 days before Ministral 3 14B-25-12, so benchmark comparisons should account for the intervening progress.

Ministral 3 14B-25-12 has 14B parameters, while Qwen2.5-Math has 1.5B/7B/72B. Context windows are 256k (Ministral 3 14B-25-12) vs 4k (Qwen2.5-Math).

On MATH, Ministral 3 14B-25-12 leads at 90.4% vs Qwen2.5-Math at 85.9%.