Regenerate 2026-09-24 (2): best-speed fix — 14 compatibility speeds

After the API's speed-ranking fix (Leavitskiy/local-creator-back#174): compatibility.speed_value/speed_unit now break confidence ties by speed, with time units lower-is-better. 14 rows change (e.g. qwen-image-2-1 on RTX 5060 Ti 41.92 s -> 21.39 s; 5 LLM rows from prefill to generation tokens/s). Nothing else changes.

README.md

@@ -39,7 +39,7 @@ and 27 cards, plus 1 005 full setup guides
 This is the machine-readable form of [smeltcore.com](https://smeltcore.com). Every row carries a
 `url` back to the page it came from.
 
-Generated 2026-09-24T08:45:55+00:00 from the public read API (`https://api.smeltcore.com/api/v1`) — no private data, no
+Generated 2026-09-24T11:18:41+00:00 from the public read API (`https://api.smeltcore.com/api/v1`) — no private data, no
 credentials, reproducible by anyone.
 
 ## Configs
@@ -125,7 +125,7 @@ extend to the measurements themselves, which belong to the sites named in
   title  = {Self-Hosted AI — GPU Compatibility, Recipes and Catalogue},
   author = {smeltcore},
   url    = {https://smeltcore.com},
-  note   = {Generated 2026-09-24T08:45:55+00:00},
+  note   = {Generated 2026-09-24T11:18:41+00:00},
   year   = {2026}
 }
 ```

data/compatibility.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:97e6dff78c069cfda07b7b2b6d466e8fc63372c9884f43b536163916b7e6142f
-size 44718
+oid sha256:9809ccfe46da913e9e0da6c93dd87325aab30d31a1aae646f8f2f2fafd721ebd
+size 44719