diff --git a/backend/cpp/llama-cpp-localai-paged/docs/GB10_PARITY_PHASE0_RESULTS.md b/backend/cpp/llama-cpp-localai-paged/docs/GB10_PARITY_PHASE0_RESULTS.md
new file mode 100644
index 000000000..3585d7444
--- /dev/null
+++ b/backend/cpp/llama-cpp-localai-paged/docs/GB10_PARITY_PHASE0_RESULTS.md
@@ -0,0 +1,23 @@
+# GB10 Parity Phase 0 Results
+
+Status: in progress.
+
+## Preflight
+
+- DGX host: `promaxgb10-4ad8`
+- Docker containers: `none`
+- GPU compute apps: `none`
+- GPU lock owner: `FREE released-by-claude-fp4norm-profile 1782828229`
+- LocalAI worktree SHA: `d288a0300f36f7c126d62d997809bb03f297a3ac`
+- Local llama.cpp fork SHA: `51168c5eee2e35348d9006f0b2fab3dc6e7c01cc`
+- DGX artifact directory: `~/bench/reopen_phase0`
+
+## Baseline Runs
+
+No baseline runs have been started yet.
+
+## Open Items
+
+- Capture clean source provenance.
+- Reproduce paged prefill and decode baselines.
+- Find or recreate vLLM graph-node-traced difference-method decode artifacts.