diff --git a/.github/workflows/_checks.yml b/.github/workflows/_checks.yml
index 2ed9a4d1..71026947 100644
--- a/.github/workflows/_checks.yml
+++ b/.github/workflows/_checks.yml
@@ -29,12 +29,9 @@ jobs:
           - "3.13"
           - "3.14"
           - "3.14t"
+          - "3.15"
+          - "3.15t"
         prerelease: [false]
-        include:
-          - python-version: "3.15"
-            prerelease: true
-          - python-version: "3.15t"
-            prerelease: true
     steps:
       - uses: actions/checkout@v6
       - uses: extractions/setup-just@v4
diff --git a/docs/dev/contributing.md b/docs/dev/contributing.md
index 717693a4..19145a85 100644
--- a/docs/dev/contributing.md
+++ b/docs/dev/contributing.md
@@ -46,7 +46,7 @@ The writing rules are in `docs/agents/docs-style.md`.
 ## What CI runs
 
 Every pull request runs `just lint-ci` and `just adr-check`, `just test-ci` and `just test-docs` on
-Python 3.11 to 3.14 (3.15 as a pre-release), `just test-race --count=50` on the free-threaded
+Python 3.11 to 3.15, `just test-race --count=50` on the free-threaded
 builds, `just test-docs-integrations`, `just docs-build`, and an offline link check of every
 Markdown file.
 
diff --git a/docs/introduction/design-decisions.md b/docs/introduction/design-decisions.md
index c82e5892..98f2c57c 100644
--- a/docs/introduction/design-decisions.md
+++ b/docs/introduction/design-decisions.md
@@ -27,7 +27,7 @@ What modern-di guarantees:
   guards `register` and `add_providers` with its own lock and iterates over a snapshot of the
   provider dict, so it never gets corrupted or raises "dict changed size during iteration".
 - Free-threaded CPython (PEP 703) is supported at `2 - Beta` and tested under real multithreading
-  on the `3.14t` build. It is Beta because modern-di relies on object-publication ordering (a
+  on the `3.14t` and `3.15t` builds. It is Beta because modern-di relies on object-publication ordering (a
   reader that sees a stored reference sees fully initialized fields), and CPython implements that
   without a published memory model to guarantee it. Per-operation latency is competitive, but
   throughput does not scale across cores: atomic reference counting of the objects every resolve
diff --git a/pyproject.toml b/pyproject.toml
index e5581fa5..b1f30e10 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -14,6 +14,7 @@ classifiers = [
     "Programming Language :: Python :: 3.12",
     "Programming Language :: Python :: 3.13",
     "Programming Language :: Python :: 3.14",
+    "Programming Language :: Python :: 3.15",
     "Programming Language :: Python :: Free Threading :: 2 - Beta",
     "Typing :: Typed",
     "Topic :: Software Development :: Libraries",