ModelCap
Current AI model comparison
Exact identities · public evidence
Qwen
Qwen3.5-9B
Position
#107
Index
41.3
Estimated
publisher-corpus-prior over 186 held-out anchors (29%); launch card against 3 resolved peers on 10 rows (71%); exceeds every named peer on 4 of 10 rows; optimism haircut 0.6 from cross-lab-probe; shrunk 3.6 up toward the measured corpus
Comparison side 1
Tencent
ContextPilot E4B
Position
#106
Index
41.3
Inherited
finetune of google/gemma-4-E4B-it · finetune (100%)
Comparison side 2
Output / 1M
$0.15
vs Unavailable
Context
262K
vs —
Weights
Open weights
vs Restricted license