diff --git a/lua/explain-it/config.lua b/lua/explain-it/config.lua
index e513440..027add4 100644
--- a/lua/explain-it/config.lua
+++ b/lua/explain-it/config.lua
@@ -8,7 +8,7 @@ M.options = {
   debug = false,
   max_notification_width = 200,
   max_retries = 3,
-  openai_chat_model = "gpt-5",
+  openai_chat_model = "gpt-5.4-codex",
   openai_completion_model = "text-davinci-003",
   output_directory = "/tmp/explain_it_output",
   split_responses = true,
diff --git a/progress.md b/progress.md
new file mode 100644
index 0000000..ccd2c45
--- /dev/null
+++ b/progress.md
@@ -0,0 +1,19 @@
+# Progress
+
+## Step: Explore repository structure
+Status: complete
+What was done: Explored the neovim plugin codebase to locate model configuration files.
+Key findings: Default model is set in `lua/explain-it/config.lua` as `openai_chat_model`. Tests use a fake model override so default model changes don't affect tests.
+
+## Step: Update default model to gpt-5.4-codex
+Status: complete
+What was done: Changed `openai_chat_model` from `"gpt-5"` to `"gpt-5.4-codex"` in `lua/explain-it/config.lua`.
+Key findings: Test run command is `make test` (requires `make deps` first if deps folder doesn't exist).
+
+## Step: Run tests
+Status: complete
+What was done: Ran `make deps && make test` — all 38 tests pass across 7 test files.
+
+## Step: Commit and open PR
+Status: complete
+What was done: Committed change with message "chore: Add gpt-5.4-codex as default model", pushed branch, and opened PR #18 at https://github.com/tdfacer/explain-it.nvim/pull/18.