ModelCap

Current AI model comparison

Exact identities · public evidence

Cohere

North Mini Code

Position

#98

Index

49.1

Measured

1 public benchmark observation across 1 board

Comparison side 1

StepFun

Step 3.7 Flash

Position

#97

Index

49.2

Estimated

3 reported rows against measured corpus ladders (17%); global-corpus-prior over 195 held-out anchors (18%); measured predecessor stepfun/step-3.5-flash less the succession penalty (65%); shrunk 0.8 up toward the measured corpus

Comparison side 2

Output / 1M

Free

vs $1.15

Context

256K

vs 262K

Weights

Open weights

vs Open weights