ModelCap
Current AI model comparison
Exact identities · public evidence
Microsoft
Phi 3.5 MoE instruct
Position
#162
Index
32.0
Estimated
global-corpus-prior over 203 held-out anchors (42%); launch card against 4 resolved peers on 6 rows (58%); exceeds every named peer on 4 of 6 rows; optimism haircut 0.67 from cross-lab-probe; shrunk 16.8 up toward the measured corpus
Comparison side 1
Microsoft
UserLM 8b
Position
#238
Index
7.1
Inherited
finetune of meta-llama/Llama-3.1-8B · finetune (100%)
Comparison side 2
Output / 1M
Unavailable
vs Unavailable
Context
—
vs —
Weights
Open weights
vs Open weights