Home › Evidence › Records › cfg-0026

cfg-0026

backfilled
citable URL: https://halobench.com/records/cfg-0026/ — this address never moves; the anchor /records/#cfg-0026 keeps resolving
modelQwen3.5-122B-A10B-UD-Q4_K_M-00001-of-00003.gguf · UD-Q4_K_M
engineggml-org/llama.cpp 3653e6d6d547ec763317d9ecd0ace334a7e21359 · rocm · host aihydra (igpu)
flags-ngl 999 -fa on -c 32768 --parallel 1 --load-mode none --jinja -rea off -ctk q8_0 -ctv q8_0
templatenot recorded at test time
treeupstream — stock
aged evidence — reconstructed from the archive. cfg-0025 with quantised KV — the arms differ ONLY in -ctk/-ctv, which is the whole design of the clm-0038 paired comparison. Same provenance and gaps as cfg-0025.

Capability basis

measured on this config

Runs on this config (1)

rundatekindsuitemetrics
run-01012026-08-10capabilitytau2-bench-airline@668d3bctasks_passed=8 · tasks_total=9 · mean_score=0.8889