Jamba2 Mini 52B: benchmark results

Provider: AI21 Labs. Access: API.

Aggregate Rating 1490 ± 28, rank #2513 of 10916 rated models, from 28 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Medmarks - MedHallu Hard53.44Score (%)85.2
Medmarks - MedHallu Easy68.99Score (%)68.2
Medmarks - MedHallu Medium61.18Score (%)61.4
Medmarks - PubMedQA73.87Score (%)42
Medmarks - MedConceptsQA Medium49.1Score (%)36.4
Medmarks - MedConceptsQA Hard48.3Score (%)31.8
Medmarks - PubHealthBench Reviewed82.15Score (%)28.4
Medmarks - MedCalc-Bench32.45Score (%)23.9
Medmarks - MedConceptsQA Easy76.27Score (%)22.7
Medmarks - CareQA EN78.77Score (%)20.5
Medmarks - LongHealth Task 182.08Score (%)20.5
Medmarks - MedMCQA59.79Score (%)20.5

Interactive version: theaggregate.ai/model?slug=jamba2-mini-52b · How It Works · Data refreshed daily, snapshot 2026-11-08.