I benchmarked "Jev-killer" OpenAI Decisions API against Jev

Back to Devoured - October 08, 2026

DEVOURED

I benchmarked "Jev-killer" OpenAI Decisions API against Jev

HiringCafe founder Hamed Nilforoshan found his custom model, Jev, outperformed OpenAI's new Decisions API and Gemini 3.1 Flash-Lite in job matching relevance.

What: Hamed Nilforoshan benchmarked his internal model, Jev, against OpenAI’s Decisions API and Google’s Gemini 3.1 Flash-Lite. Jev achieved higher Spearman correlation scores in matching user queries and resumes to job descriptions while reportedly costing less.

Why it matters: The rise of specialized, task-specific models suggests that general-purpose 'decisions' APIs from major AI labs may struggle to compete with models tuned specifically for narrow, high-frequency enterprise workflows.

Decoder

Spearman Correlation: A statistical measure of the strength and direction of the association between two ranked variables; in this context, it measures how well the AI's relevance ranking matches human-labeled relevance.

Original article

I benchmarked "Jev-killer" OpenAI Decisions API against Jev for HiringCafe.com, serving 2.5 million users. The task is to predict how relevant a user query/resume is to a job description, on a scale of 1-10. OpenAI is 2x more expensive and 5-10% worse.

Task 1: Relevance user query x job description

Jev: Spearman Corr. 0.74
OpenAI: Spearman Corr. 0.71
Gemini 3.1 FL: Spearman Corr. 0.63

Task 2: Relevance of a resume x job description

Jev: Spearman Corr. 0.71
OpenAI: Spearman Corr. 0.67
Gemini 3.1 FL: Spearman Corr. 0.61

Check out our AI job search at We use Jev to match your resume against (almost) every job on earth!! Totally free (and yes we're burning a lot on Jev tokens, its our #1 biggest expense). hiringcafe.com

Explore all 59 articles in today's digest