llama.cpp ROCm AI Max+ 395 llama-2-7b Q4_0 tg128 FA
49.87 tok/s (tg128)
Fields
id
benchmarks/llamacpp-scoreboard-ryzen-ai-max-395-framework-desktop-fa-tg128Metric
decode_tok_s
Value
49.87
Unit
tok/s (tg128)
Model
llama-2-7b.Q4_0.gguf (ggml-org, 3.56 GiB, 6.74B params)
Methodology
COMMUNITY-SCOREBOARD (llama.cpp discussion #15021, maintainer-curated ROCm backend). llama-bench llama-2-7b.Q4_0.gguf (ggml-org, 3.56 GiB, 6.74B params), pp512 and tg128, n_gpu_layers 99, batch 1, full GPU offload, ROCm. Per-request batch-1, not an aggregate across concurrent requests. Flash attention on (-fa 1). Build e60f241, reporter @firefox42. Printed mean 49.87. Re-read of the discussion page on 2026-10-17.
Measured by
third_party
Reproducible
yes
notes
AI Max+ 395, 128 GB / LPDDR5, build e60f241, reporter @firefox42. token generation cell.
Notes
AI Max+ 395, 128 GB / LPDDR5, build e60f241, reporter @firefox42. token generation cell.