Qwen3.6-35B-A3B Q4_K_M on Ryzen AI Max+ 395 (llama.cpp Vulkan) — answerless (empty) responses at a 1k budget: 75.0 % of requests

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 answerless (empty) responses at a 1k budget was 75.0 % of requests (share of all issued requests). The measurement was taken under the frozen interactive-assistant-v1@2026-08-02 workload; evidence level lab_repeated, 3 valid runs across 1 physical unit. The value is re-derived from the raw run records on every CI build.

Statement

At a 1024-token budget with reasoning enabled, this share of requests returned HTTP 200 with an empty answer: the reasoning pass consumed the entire budget.

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

interactive-assistant-v1@2026-08-02

Aggregation

share of all issued requests

Evidence level

lab_repeated

Status

active

Limitations

Three runs of 16 requests each (48 total) on one unit — repeated, not unit-replicated. Specific to a 1024-token budget and this corpus. Failed requests are counted in the denominator, as required.

Runs

• run-20260802-gonogo-a2
• run-20260802-budget1k-a-r2
• run-20260802-budget1k-a-r3

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/answerless-rate-1k.sql

Cite this claim

AGmind Systems Lab. "At a 1024-token budget with reasoning enabled, this share of requests returned HTTP 200 with an empty answer: the reasoning pass consumed the entire budget." Claim strix.qwen36.interactive.c1.answerless-1k (75.0 % of requests), evidence level lab_repeated, scope interactive-assistant-v1@2026-08-02. https://agmind.ai/claims/strix.qwen36.interactive.c1.answerless-1k/

@misc{agmind_strix_qwen36_interactive_c1_answerless_1k,
  author       = {{AGmind Systems Lab}},
  title        = {At a 1024-token budget with reasoning enabled, this share of requests returned HTTP 200 with an empty answer: the reasoning pass consumed the entire budget.},
  howpublished = {\url{https://agmind.ai/claims/strix.qwen36.interactive.c1.answerless-1k/}},
  note         = {Claim strix.qwen36.interactive.c1.answerless-1k: 75.0 \% of requests; evidence level lab\_repeated; scope interactive-assistant-v1@2026-08-02},
  year         = {2026}
}

Machine-readable: /claims/strix.qwen36.interactive.c1.answerless-1k.json · full registry