ModelCap

Current AI model comparison

Exact identities · public evidence

Meta

Llama 4 Maverick

Position

#99

Index

25.4

Measured

4 public benchmark observations across 4 boards

Comparison side 1

Qwen

Qwen2.5 7B Instruct

Position

#102

Index

24.1

Estimated

publisher corpus prior · leave-one-anchor-out calibrated

Comparison side 2

Output / 1M

$0.80

vs $0.20

Context

1M

vs 33K

Weights

Gated access

vs Open weights