# Qwen-Flast-Next (UD-IQ4_XS v3.1) running 148-182 tok/s (1.6-1.8x strata) while KLD MEAN = 0.08195 — retained record

> Retained reddit record from reddit:LocalLLaMA.

Canonical: https://highsignal.app/data/records/e61c197365757fe4

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:LocalLLaMA)  published 2026-11-17 19:53 ingested 2026-11-18 02:42

[original source ↗](https://www.reddit.com/r/LocalLLaMA/comments/1xyue7j/qwenflastnext_udiq4_xs_v31_running_148182_toks/)

## retained text



```
Numbers only, rest is in the repo.

# Quality against BF16

|KV|PPL|mean KLD|top-1|paired dKLD vs int8 \[95% CI\]|prefill 131K|
|:-|:-|:-|:-|:-|:-|
|**int8**|3.2323|**0.08195**|**90.45%**|reference|**6,019 t/s**|
|fp16|3.2320|0.08230|90.33%|\+0.00035 \[-0.00061, +0.00131\]|2,997 t/s|
|k8v4|3.2424|0.08621|90.26%|\+0.00426 \[+0.00307, +0.00545\]|5,694 t/s|
|q4\_0|3.2399|0.08843|90.03%|\+0.00647 \[+0.00499, +0.00795\]|5,700 t/s|

# Speed

|Test|Halocline|Strata 0.2.10, same card and quant|Ratio|
|:-|:-|:-|:-|
|Warm decode, short prompt|148 to 182 tok/s|95 to 101 tok/s|1.6 to 1.8x|
|36K needle: decode / prompt|123 / 6,298 tok/s|68 / 2,403 tok/s|1.8x / 2.6x|
|137K needle: decode / prompt|121 / 5,664 tok/s|87 / 2,864 tok/s|1.4x / 2.0x|
```



document fetched 2026-11-18 00:45 · canonical https://reddit.com/r/LocalLLaMA/comments/1xyue7j/qwenflastnext_udiq4_xs_v31_running_148182_toks
