M3 Max llama-3-1-8b-instruct Q4_K - Medium decode (llamafile, siliconscore ref)
37.47979072222222 tok/s (single-request decode, per-request)
Fields
id
benchmarks/m3max-llama-3-1-8b-instruct-q4-k-medium-llamafile-ctxnp-tgMetric
decode_tok_s
Value
37.4798
Unit
tok/s (single-request decode, per-request)
Model
llama-3-1-8b-instruct
Methodology
Silicon Score benchmarks.csv reference run. M3 Max (30-core GPU, 36 GB): llama-3-1-8b-instruct Q4_K - Medium via llamafile. context_tokens=not published. Decode tok/s single-request.
Measured by
third_party
Reproducible
yes
Thermal state
unknown
notes
siliconscore reference run; source_url=https://www.localscore.ai/accelerator/805. Chip variant 'M3 Max (30-core GPU, 36 GB)'.
Notes
siliconscore reference run; source_url=https://www.localscore.ai/accelerator/805. Chip variant 'M3 Max (30-core GPU, 36 GB)'.