30GB smaller, and higher quality
Our ROCmFPx build of Qwen3.8-Flash-Next is 30GB smaller than the leading community IQ4_XS quant, and it comes out ahead on perplexity too, not behind.
August 28, 2026
Size and quality are normally a trade
Quantizing a model smaller almost always costs you something in quality. The usual question is just how much you're willing to pay. ROCmFPx, the 4-bit floating-point format we built for Strix Halo's unified memory (see the adaptive speculation writeup for the rest of that fork), was built to bend that trade rather than just ride it.
We compared our ROCmFP4-FAST-imatrix build of Qwen3.8-Flash-Next against the leading community IQ4_XS requant of the same model, from quantizer AesSedai.
Smaller, not worse
| Build | Size | Δ perplexity vs. reference |
|---|---|---|
| AesSedai IQ4_XS | 117.13 GiB | +3.12% |
| Agention ROCmFP4-FAST-imatrix | 87.06 GiB | +2.48% |
30.07 GiB smaller, with a smaller perplexity penalty too: +2.48% versus +3.12% relative to the unquantized reference. A quant this much smaller usually costs quality; this one doesn't.
Get it
The build is sized for Strix Halo's 96GB unified memory and runs entirely on GPU. Full benchmark methodology and download on HuggingFace, or see it alongside our other builds on the models page.
View on HuggingFace ↗