ModelCap
Current AI model comparison
Exact identities · public evidence
Qwen
Qwen3.5 4B
Position
#120
Index
38.3
Estimated
publisher-corpus-prior over 187 held-out anchors (26%); launch card against 3 resolved peers on 10 rows (74%); exceeds every named peer on 2 of 10 rows; optimism haircut 0.6 from cross-lab-probe; shrunk 3.9 up toward the measured corpus
Comparison side 1
Tiiuae
Falcon H1 34B Instruct
Position
#119
Index
38.8
Estimated
global-corpus-prior over 187 held-out anchors (46%); launch card against 5 resolved peers on 11 rows (54%); exceeds every named peer on 3 of 11 rows; optimism haircut 0 from cross-lab-probe; shrunk 11.1 up toward the measured corpus
Comparison side 2
Output / 1M
Unavailable
vs Unavailable
Context
—
vs —
Weights
Open weights
vs Restricted license