diff --git a/README.md b/README.md
index a53117d..0ae921f 100644
--- a/README.md
+++ b/README.md
@@ -95,6 +95,20 @@ differences. That run compares the two engines; that the location rows answer
 as the reference did, and that the statement programs compile as it compiled
 them, is what the suite checks on the server runtime.
 
+Run again on 2026-10-04, on this source as it stood just before the version
+moved to 0.6.0, at `9a2fb80`: the source of the run above plus the two source
+changes made after it, the fraction expander's bound computed from the unit
+table and the exported name for `evaluateTagged`'s result type. The same four
+reports on the same engine build (release `0.12.0`, bytecode version `89`): the
+evaluator surface answered 257 rows on the server runtime and 257 on the
+engine, the compiler surface 215 and 215, the location run 112 and 112 and the
+statement run 64 and 64, all four with zero differences. The corpus is the
+same tier 9, 262 cases, as in the runs above. What it covers is what the four
+reports run: the corpus cases as the 2026-10-02 paragraph describes them, the
+three location functions with each transcript row's inputs, and the statement
+programs through `executeTagged`. Like every run here, it is a run on the
+standalone build named above, not on the engine an application ships.
+
 `engines.node` in `package.json` is `>=20`, and that is the floor a
 consumer's runtime has to clear. It is not the toolchain: what builds and
 gates this repository is the one node and the one pnpm `mise.toml` pins, and
diff --git a/conformance/README.md b/conformance/README.md
index b4c039c..68c73b1 100644
--- a/conformance/README.md
+++ b/conformance/README.md
@@ -206,6 +206,25 @@ The proof it produces is against the standalone command-line build of that
 engine, which is an older release than the one current React Native ships, and
 the script's header says what that bounds.
 
+### The run before 0.6.0 (2026-10-04)
+
+Run on 2026-10-04 on the source at `9a2fb80`, as it stood just before the
+version moved to 0.6.0, with the VM at Hermes release 0.12.0, bytecode version
+89, and the server runtime at Node v24.21.0. The corpus was tier 9, 262 cases;
+beside it ran the 112 location transcript rows and the 64 statement programs.
+It printed:
+
+| Report | Server runtime | The VM | Differences |
+|---|---|---|---|
+| evaluator | 257 rows | 257 rows | 0 |
+| compiler | 215 rows | 215 rows | 0 |
+| location | 112 rows | 112 rows | 0 |
+| statement | 64 rows | 64 rows | 0 |
+
+The README records this run beside the earlier ones: the 2026-09-20 and
+2026-10-02 runs of the two corpus surfaces, and the first run of all four
+reports, at `c901cc0`. Every one of them diffed clean.
+
 ### A loop gives every closure one binding on that engine (2026-10-04)
 
 The standalone build does not give each iteration of a loop its own binding