gpt-oss-120bbenchedguard failed
②Verdict
Excluded. The guard failed: the model returns stale answers from previous requests — the first request on a fresh server answers correctly and every subsequent request returns the first one's answer, reproduced 6/6 at --parallel 1 with prompt caching ruled out clm-0025. A model that occasionally serves a previous answer is far worse than a slow one, so it is not a candidate for any role until the staleness is understood. Its throughput rows stay recorded but not guard-cleared: the fastest decode at empty context in the field, 55.45 tok/s, and those numbers say nothing about whether it generates the right thing clm-0025. Flash attention is mandatory — both fa-off arms failed outright — and q8_0 KV costs 56% of decode at 131k on the stock build, more than anything else measured clm-0025. On the blind eight-task conceptual screen it placed last at 3/8; that result is one date, one quant, one build, N=1 per cell, and is not a general capability verdict clm-0004.
③Best configuration
| model | gpt-oss-120b-UD-Q4_K_XL-00001-of-00002.gguf · UD-Q4_K_XL |
| engine | ggml-org/llama.cpp 3653e6d · rocm · host aihydra (igpu) |
| flags | -ngl 999 -fa 1 -b 2048 -ub 512 -ctk f16 -ctv f16 -t 16 |
| template | not recorded at test time |
| tree | upstream — stock |
backfilled aged evidence — reconstructed from the archive · config record cfg-0016
every cell generated from the record at build time · throughput cells from cfg-0016 (same build 3653e6d, fa on, f16 KV) · provenance: measured-here throughout
③aDecode against context depth
④Other configurations tested — each as a delta against best
| variant | Δ decode | Δ turns (paired tasks) | Δ energy | note | records |
|---|---|---|---|---|---|
| q8_0 KV · stock build | -56% @131k | — | — | The steepest quantised-KV penalty in the field — which also makes this model the largest potential beneficiary of the KV-dequant patch, if it is ever cleared to matter. | clm-0025 clm-0022 |
| flash attention off hazard | — | failed outright — no series exists | Both flash-attention-off arms (f16 and q8_0 KV) failed to produce a run — recorded as failures against their fingerprints, not retry candidates. | clm-0025 | |
deltas computed at build time from the named runs, paired-task traces and metered windows — a hazard row shows words where a delta would mislead
⑤Open questions
- The staleness mechanism — slot state not reset between requests is the pattern; the harmony chat-format code path is the hypothesis, unconfirmed. Three-line reproduction against a fresh server; upstream report owed once characterised clm-0025.
- The KV-dequant patch on this model — the largest measured q8_0 penalty in the field and no patched arm run clm-0025 clm-0022.
- The 3.5% CV flag on the f16 cell at 131k — above the scatter threshold, a defective-path signature warranting a look rather than a shrug clm-0025 clm-0017.