ModelCap

Current AI model comparison

Exact identities · public evidence

Microsoft

MAI DS R1

Position

#95

Index

46.8

Inherited

finetune of deepseek/deepseek-r1 · finetune (100%)

Comparison side 1

StepFun

Step 3.7 Flash

Position

#94

Index

46.9

Estimated

3 reported rows against measured corpus ladders (11%); global-corpus-prior over 187 held-out anchors (21%); measured predecessor stepfun/step-3.5-flash less the succession penalty (68%); shrunk 1.3 up toward the measured corpus

Comparison side 2

Output / 1M

Unavailable

vs $1.15

Context

vs 262K

Weights

Open weights

vs Open weights