Qwen3.6-35B-A3B Q4_K_M on Ryzen AI Max+ 395 (llama.cpp Vulkan) — decode-pace drift over three hours, 4 concurrent requests: 1.6 %

The measurement

On a Beelink GTR9 Pro — AMD Ryzen AI Max+ 395, Radeon 8060S (gfx1151), 128 GB LPDDR5X-8000 unified running llama.cpp (server, Vulkan backend), b9049 (server_fingerprint b9049-2496f9c14) with ggml-org/Qwen3.6-35B-A3B-GGUF @ baec3ebee244 (Q4_K_M), the measured decode-pace drift over three hours, 4 concurrent requests was 1.6 % (last window vs first window, pooled over both units). The measurement was taken under the frozen endurance-30m-v1@2026-08-03 workload; evidence level lab_single_run, 2 valid runs across 2 physical units. The value is re-derived from the raw run records on every CI build.

Statement

Relative change of the median inter-token latency between the first five minutes and minutes 175-180 of a continuous 3-hour closed-loop pass, across both units.

Evidence

System

Beelink GTR9 Pro — AMD Ryzen AI Max+ 395, Radeon 8060S (gfx1151), 128 GB LPDDR5X-8000 unified

Runtime

llama.cpp (server, Vulkan backend), b9049 (server_fingerprint b9049-2496f9c14)

Model artifact

ggml-org/Qwen3.6-35B-A3B-GGUF @ baec3ebee244 (Q4_K_M)

Scope

endurance-30m-v1@2026-08-03

Aggregation

last window vs first window, pooled over both units

Evidence level

lab_single_run

Status

active

Published

August 3, 2026

Limitations

One 3-hour pass per unit — the pattern reproduced on both commercially identical units, but repeats within a unit are pending. Closed-loop concurrency 4, reasoning disabled, this corpus — not a general thermals verdict. Ambient not instrumented; die temperatures documented in the run notes.

Runs

• run-20260803-end-c4-a
• run-20260803-end-c4-b

Derivation

The value is re-derived from the raw run records by this query on every CI build — a published number cannot silently drift from its evidence.

catalog/claims/sql/itl-drift-180m.sql

Cited on

Pages on this site that render this number:

Cite this claim

AGmind Systems Lab. Qwen3.6-35B-A3B Q4_K_M on Ryzen AI Max+ 395 (llama.cpp Vulkan): decode-pace drift over three hours, 4 concurrent requests — 1.6 % (last window vs first window, pooled over both units; 2 runs on 2 units; evidence level lab_single_run; workload endurance-30m-v1@2026-08-03). Claim strix.qwen36.endurance.c4.itl-drift-180m. https://agmind.ai/claims/strix.qwen36.endurance.c4.itl-drift-180m/

@misc{agmind_strix_qwen36_endurance_c4_itl_drift_180m,
  author       = {{AGmind Systems Lab}},
  title        = {Relative change of the median inter-token latency between the first five minutes and minutes 175-180 of a continuous 3-hour closed-loop pass, across both units.},
  howpublished = {\url{https://agmind.ai/claims/strix.qwen36.endurance.c4.itl-drift-180m/}},
  note         = {Claim strix.qwen36.endurance.c4.itl-drift-180m: 1.6 \%; evidence level lab\_single\_run; scope endurance-30m-v1@2026-08-03},
  year         = {2026}
}

Machine-readable: /claims/strix.qwen36.endurance.c4.itl-drift-180m.json · BibTeX · CSL-JSON · full registry · changes feed