Decode · tok/s
53.8
Prefill
1293
tok/100W
38.4
Measured on a Ryzen AI Max+ 395 with 128 GB, UD-IQ4_XS, medians of interleaved runs: 8K 1,293 prompt / 53.8 output.
Vendor release table, not a street rerun. Rank is the quoted 53.8 at 8K. 46.6 and 51.4 stay in the note.
