Model card: record the verified Hub round-trip (fresh download, CPU replay)

README.md

@@ -94,10 +94,13 @@ forecasts `p[1,12,4,40701]` (SSTA, K), the causal context `context[1,3,40701]` (
 observed SSTA) and `valid_dates[1,12]`; it contains no target values. See the package [README](frozen_model_20261025_v1/README.md)
 for the full contract.
 
-**Verified.** On CPU the prepared-input path reproduces the release arrays exactly (max |Δ| = 0 for mean, σ and CDF;
-`CPU_SMOKE_TEST.json`). The raw-input path was run on the original workstation GPU when the release was made (max SST difference
-6.1 × 10⁻⁵ K against the frozen export, float32 quantisation); the raw GPU forward was not re-run while freezing this
-package. On other hardware or library versions results are not guaranteed to be bitwise identical.
+**Verified.** A fresh download of tag `release-20261025` passes `--verify-only` (59 files), and the CPU replay of
+`examples/input_only_202610.npz` reproduces `forecasts/sst_forecast_202610_202709_physical_grid.nc` to 6.1 × 10⁻⁵ K for
+SST and quantiles, 3.5 × 10⁻⁵ K for SSTA, 4 × 10⁻⁷ K for σ and 1.1 × 10⁻⁴ for the quintile probabilities, with an
+identical NaN pattern and identical thresholds. That is the documented float32 difference between the prepared-input
+path and the raw-input GPU path that produced the forecast; against the original prepared-input API the CPU path is
+exact (`CPU_SMOKE_TEST.json`). The raw GPU forward was not re-run while freezing this package. On other hardware or
+library versions results are not guaranteed to be bitwise identical.
 
 ## Latest forecast