docs: update jerboa-db.md header with Phase 5 status

ober

800d72ff99a4b1798ac01951c4464b441a43094e

diff --git a/jerboa-db.md b/jerboa-db.md
index a4c7ef1..b96dd43 100644
--- a/jerboa-db.md
+++ b/jerboa-db.md
@@ -1,19 +1,22 @@
 # Jerboa-DB: Datomic Without the JVM
 
 **Goal:** A fully-featured Datomic clone built entirely on Jerboa, using LevelDB for
-persistent index storage and DuckDB for analytics (planned).  Single binary,
-embeddable, with a Datalog query engine and immutable time-travel over all data.
-
-**Status:** 2026-04-24 — **Core Datomic parity: complete.** Core test
-suite is 37/37 passing (`make test`), including `not-join`, the
-`variance`/`stddev` family, stored `:db/fn`, `log` API object,
-`index-range`, and `seek-datoms` — all previously tracked as gaps.
-Phase 2 (LevelDB persistence) is production-ready.  Phase 4 (DuckDB
-SQL-over-datoms) is functional with Parquet and CSV bulk I/O wired
-to DuckDB's native `COPY TO`/`COPY FROM`.  Schema migration
-(rename, retype, delete) and additive evolution all work via
-`migrate!`.  Remaining gaps: TLS transport (libssl integration
-landed in Round 8 as opt-in), CLI entrypoint.
+persistent index storage and DuckDB for analytics.  Single binary, embeddable,
+with a Datalog query engine and immutable time-travel over all data.
+
+**Status:** 2026-05-03 — **Core Datomic parity: complete.  MBrainz
+performance parity: shipped.** Core test suite is 37/37 passing
+(`make test`).  Five-phase performance plan landed between
+commits `439b9a7..7a45249`: pure-aggregate pushdown, DuckDB
+analytical fallback (with batched VALUES sync, `:in` substitution,
+auto-routing via `current-analytics-engine`), and the load-time
+fix that took 1% MBrainz ingest from ~47s to ~1s (a fulltext
+indexing O(n²) bug, not the originally-suspected RB-tree
+allocation pressure).  At 1% scale the 8-query suite finishes in
+under 3 seconds total; Q4 lands in 281 ms via DuckDB and Q8 in 9
+ms — both faster than published Datahike numbers.  See
+[MBrainz Benchmark Results](#mbrainz-benchmark-results) for the
+table.
 
 Implementation lives in `lib/jerboa-db/` (Jerboa library files using `#!chezscheme`
 + `(library ...)` form).  Built from scratch — no `(std mvcc)`, `(std datalog)`, or