Update README.md

README.md

@@ -102,7 +102,7 @@ Two things set v3 apart from earlier releases of The Stack:
 1. **The file contents are included inline.** The decoded UTF-8 source text is embedded directly in the dataset, so it is fully self-contained — you can start training the moment the download finishes.
 2. **It reflects the state of GitHub in August 2025.** The corpus is a direct crawl of GitHub repositories at their default-branch HEAD, capturing roughly two additional years of open-source code compared to The Stack v2.
 
-> The Stack v3 training subset contains **15.9 TB** of source code across **713** programming languages from **173M** repositories (~**4.9 trillion** tokens).
+> The Stack v3 training subset contains **15.9 TB** of source code across **713** programming languages from **173M** repositories (~**3.6 trillion** tokens).
 > The Full Stack v3 corpus contains **113.7 TB** of source code across **770** languages from **224M** repositories.
 
 ### The Stack across versions
@@ -112,7 +112,7 @@ Two things set v3 apart from earlier releases of The Stack:
 | cutoff | 2022 | 2023 | 2025         |
 | full | 6.4 TB | 67.5 TB | 113.7 TB     |
 | dedup | 2.9 TB | 32.1 TB | 99.3 TB      |
-| train | ~200B tokens | 2.0 TB / ~550B tokens | 15.9 TB / ~4.9T tokens |
+| train | ~200B tokens | 2.0 TB / ~550B tokens | 11.5 TB / ~3.6T tokens |
 | languages | 358 | 618 | 713          |
 
 ## What is being released
@@ -132,6 +132,7 @@ The Stack v3 is published in two complementary datasets:
 
 | Release | Description |
 |---|---|
+| v3.1 | Removed exact-duplicate files that leaked due to a partitioning bug, 4.9T -> 3.6T tokens. |
 | v3.0 | Initial release of The Stack v3. |
 
 ## How to download and use it