analytical-viz: capacity planning tab — add public-example B/replica table

Section 6 (new): 'Actually-provisioned batch sizes (public examples)'.
Ballpark B/replica figures for Gemini 1.5 Pro/Flash 1M, GPT-4.1 1M,
Claude 3.5 Sonnet 200k, and enterprise dedicated tiers, with the
inference basis for each ('pricing ~2x standard suggests low B',
etc). Caption flags that none are officially published; numbers are
backed out from public pricing + SLAs.

Binding-axis playbook renumbered #6 → #7.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

tests/analytical_visualization/app.py

@@ -2888,8 +2888,38 @@ with tab_planning:
 
     st.divider()
 
-    # ── Section 6: Binding-axis playbook ──────────────────────────
-    st.markdown("### 6. What to do when each axis binds")
+    # ── Section 6: Real-world provider examples ───────────────────
+    st.markdown("### 6. Actually-provisioned batch sizes (public examples)")
+    st.caption(
+        "None of these are officially published. The numbers are "
+        "**backed out from public pricing** (long-context tiers cost "
+        "2–5× per token vs standard) and stated latency SLAs. Treat "
+        "as ballpark, not authoritative."
+    )
+    _real_rows = [
+        {"Provider / tier": "Gemini 1.5 Pro (1M)",
+         "Approx B / replica": "1–4",
+         "Signal": "Pricing ~2× standard suggests low B"},
+        {"Provider / tier": "Gemini 1.5 Flash (1M)",
+         "Approx B / replica": "2–8",
+         "Signal": "Smaller model → more HBM headroom for KV"},
+        {"Provider / tier": "GPT-4.1 (1M)",
+         "Approx B / replica": "1–4",
+         "Signal": "Similar pricing structure to Gemini Pro"},
+        {"Provider / tier": "Claude 3.5 Sonnet (200k)",
+         "Approx B / replica": "4–16",
+         "Signal": "Shorter context = more headroom, higher B feasible"},
+        {"Provider / tier": "Enterprise 'dedicated' tiers",
+         "Approx B / replica": "1",
+         "Signal": "Whole replica per customer for latency guarantees"},
+    ]
+    st.dataframe(pd.DataFrame(_real_rows), width='stretch',
+                  hide_index=True)
+
+    st.divider()
+
+    # ── Section 7: Binding-axis playbook ──────────────────────────
+    st.markdown("### 7. What to do when each axis binds")
     _playbook_rows = [
         {"Binding axis": "A. Capacity (weights)",
          "Symptom": "Weights alone are close to per-PE HBM budget",