Tenstorrent Wormhole n150 Mistral-7B token-gen (vendor b32)
28.7 tok/s per user at batch-32; 918 t/s aggregate
Fields
id
benchmarks/comm-tt-wormhole-n150-mistral-7b-tgMetric
tok_s_per_user
Value
28.7
Unit
tok/s per user at batch-32; 918 t/s aggregate
Model
Mistral-7B
Methodology
Vendor tt-metal model matrix, batch-32 decode, ISL=128
Measured by
vendor
Thermal state
unknown
notes
VENDOR-MEASURED. Highest t/s/u figure on n150 in the matrix.
Notes
VENDOR-MEASURED. Highest t/s/u figure on n150 in the matrix.