ModelCap

Current AI model comparison

Exact identities · public evidence

Microsoft

Phi 3.5 mini instruct

Position

#188

Index

24.1

Estimated

global-corpus-prior over 210 held-out anchors (26%); launch card against 4 resolved peers on 6 rows (74%); exceeds every named peer on 1 of 6 rows; optimism haircut 0.67 from cross-lab-probe; shrunk 11.4 up toward the measured corpus

Comparison side 1

NVIDIA

Nemotron 3 Content Safety

Position

#191

Index

22.9

Inherited

finetune of google/gemma-3-4b-it · finetune (100%)

Comparison side 2

Output / 1M

Unavailable

vs Unavailable

Context

vs

Weights

Open weights

vs Restricted license