Home › Evidence › Records › con-0017
⚠ This notebook has stopped taking notes: newest record is 23 days old (eng-0280), against an expected cadence of 14 days.

con-0017

carrying
citable URL: https://halobench.com/records/con-0017/ — this address never moves; the anchor /records/#con-0017 keeps resolving
kind report · upstream kingjones777/Qwen3.8-Flash-Next-ROCmFP4-STRIX-GGUF
opened 2026-08-27 · https://huggingface.co/kingjones777/Qwen3.8-Flash-Next-ROCmFP4-STRIX-GGUF/blob/069dddb53bab04218d734fa9a771f8a0242ab059/README.md
carrying — reproducibility hazard: in use but not upstream. Every config listing this id in its divergence runs a fork.
Immutable publisher model-page revision 069dddb53bab04218d734fa9a771f8a0242ab059 describes the full-STRIX artifact and the required kingjones30/ROCmFPX runtime. Its long-context table was measured on gfx1151 with ROCm 7.2.4, different target prompt counts and context allocations, and no HaloBench 512-checkpoint/ 2048-token-cadence proof. The page reports full-STRIX rows through c131072, while its c262144 rows are explicitly STRIX_LEAN. HaloBench used the full-STRIX artifact on HIP 7.1.52801-9999, exact cache-false prompt_n targets, c262144, and the corrected checkpoint flags. The 32k-131k rates are close in magnitude, but neither source transfers: runtime version, prompt bytes/counts, allocation, repetition method and checkpoint boundary differ. The publisher page is a comparison source and attribution caveat, not evidence for a HaloBench number.

Cited by — computed at build time, never stored

claims clm-0127
model pages qwen38-flash-next