fix(coding-agent): make resolveCliModel sync, update docs and changelog

packages/coding-agent/CHANGELOG.md

@@ -10,6 +10,7 @@
 
 - Fixed context usage percentage in footer showing stale pre-compaction values. After compaction the footer now shows `?/200k` until the next LLM response provides accurate usage ([#1382](https://github.com/badlogic/pi-mono/pull/1382) by [@ferologics](https://github.com/ferologics))
 - Fixed `_checkCompaction()` using the first compaction entry instead of the latest, which could cause incorrect overflow detection with multiple compactions ([#1382](https://github.com/badlogic/pi-mono/pull/1382) by [@ferologics](https://github.com/ferologics))
+- `--model` now works without `--provider`, supports `provider/id` syntax, fuzzy matching, and `:<thinking>` suffix (e.g., `--model sonnet:high`, `--model openai/gpt-4o`) ([#1350](https://github.com/badlogic/pi-mono/pull/1350) by [@mitsuhiko](https://github.com/mitsuhiko))
 
 ## [0.52.9] - 2026-02-08
 

packages/coding-agent/README.md

@@ -452,7 +452,7 @@ pi config                   # Enable/disable package resources
 | Option | Description |
 |--------|-------------|
 | `--provider <name>` | Provider (anthropic, openai, google, etc.) |
-| `--model <id>` | Model ID |
+| `--model <pattern>` | Model pattern or ID (supports `provider/id` and optional `:<thinking>`) |
 | `--api-key <key>` | API key (overrides env vars) |
 | `--thinking <level>` | `off`, `minimal`, `low`, `medium`, `high`, `xhigh` |
 | `--models <patterns>` | Comma-separated patterns for Ctrl+P cycling |
@@ -524,6 +524,12 @@ pi -p "Summarize this codebase"
 # Different model
 pi --provider openai --model gpt-4o "Help me refactor"
 
+# Model with provider prefix (no --provider needed)
+pi --model openai/gpt-4o "Help me refactor"
+
+# Model with thinking level shorthand
+pi --model sonnet:high "Solve this complex problem"
+
 # Limit model cycling
 pi --models "claude-*,gpt-4o"
 

packages/coding-agent/docs/rpc.md

@@ -12,7 +12,7 @@ pi --mode rpc [options]
 
 Common options:
 - `--provider <name>`: Set the LLM provider (anthropic, openai, google, etc.)
-- `--model <id>`: Set the model ID
+- `--model <pattern>`: Model pattern or ID (supports `provider/id` and optional `:<thinking>`)
 - `--no-session`: Disable session persistence
 - `--session-dir <path>`: Custom session storage directory
 

packages/coding-agent/src/cli/args.ts

@@ -244,6 +244,12 @@ ${chalk.bold("Examples:")}
   # Use different model
   ${APP_NAME} --provider openai --model gpt-4o-mini "Help me refactor this code"
 
+  # Use model with provider prefix (no --provider needed)
+  ${APP_NAME} --model openai/gpt-4o "Help me refactor this code"
+
+  # Use model with thinking level shorthand
+  ${APP_NAME} --model sonnet:high "Solve this complex problem"
+
   # Limit model cycling to specific models
   ${APP_NAME} --models claude-sonnet,claude-haiku,gpt-4o
 

packages/coding-agent/src/core/model-resolver.ts

@@ -274,11 +274,11 @@ export interface ResolveCliModelResult {
  * Note: This does not apply the thinking level by itself, but it may *parse* and
  * return a thinking level from "<pattern>:<thinking>" so the caller can apply it.
  */
-export async function resolveCliModel(options: {
+export function resolveCliModel(options: {
 	cliProvider?: string;
 	cliModel?: string;
 	modelRegistry: ModelRegistry;
-}): Promise<ResolveCliModelResult> {
+}): ResolveCliModelResult {
 	const { cliProvider, cliModel, modelRegistry } = options;
 
 	if (!cliModel) {

packages/coding-agent/src/main.ts

@@ -403,13 +403,13 @@ async function createSessionManager(parsed: Args, cwd: string): Promise<SessionM
 	return undefined;
 }
 
-async function buildSessionOptions(
+function buildSessionOptions(
 	parsed: Args,
 	scopedModels: ScopedModel[],
 	sessionManager: SessionManager | undefined,
 	modelRegistry: ModelRegistry,
 	settingsManager: SettingsManager,
-): Promise<{ options: CreateAgentSessionOptions; cliThinkingFromModel: boolean }> {
+): { options: CreateAgentSessionOptions; cliThinkingFromModel: boolean } {
 	const options: CreateAgentSessionOptions = {};
 	let cliThinkingFromModel = false;
 
@@ -421,7 +421,7 @@ async function buildSessionOptions(
 	// - supports --provider <name> --model <pattern>
 	// - supports --model <provider>/<pattern>
 	if (parsed.model) {
-		const resolved = await resolveCliModel({
+		const resolved = resolveCliModel({
 			cliProvider: parsed.provider,
 			cliModel: parsed.model,
 			modelRegistry,
@@ -670,7 +670,7 @@ export async function main(args: string[]) {
 		sessionManager = SessionManager.open(selectedPath);
 	}
 
-	const { options: sessionOptions, cliThinkingFromModel } = await buildSessionOptions(
+	const { options: sessionOptions, cliThinkingFromModel } = buildSessionOptions(
 		parsed,
 		scopedModels,
 		sessionManager,

packages/coding-agent/test/model-resolver.test.ts

@@ -207,12 +207,12 @@ describe("parseModelPattern", () => {
 });
 
 describe("resolveCliModel", () => {
-	test("resolves --model provider/id without --provider", async () => {
+	test("resolves --model provider/id without --provider", () => {
 		const registry = {
 			getAll: () => allModels,
 		} as unknown as Parameters<typeof resolveCliModel>[0]["modelRegistry"];
 
-		const result = await resolveCliModel({
+		const result = resolveCliModel({
 			cliModel: "openai/gpt-4o",
 			modelRegistry: registry,
 		});
@@ -222,12 +222,12 @@ describe("resolveCliModel", () => {
 		expect(result.model?.id).toBe("gpt-4o");
 	});
 
-	test("resolves fuzzy patterns within an explicit provider", async () => {
+	test("resolves fuzzy patterns within an explicit provider", () => {
 		const registry = {
 			getAll: () => allModels,
 		} as unknown as Parameters<typeof resolveCliModel>[0]["modelRegistry"];
 
-		const result = await resolveCliModel({
+		const result = resolveCliModel({
 			cliProvider: "openai",
 			cliModel: "4o",
 			modelRegistry: registry,
@@ -238,12 +238,12 @@ describe("resolveCliModel", () => {
 		expect(result.model?.id).toBe("gpt-4o");
 	});
 
-	test("supports --model <pattern>:<thinking> (without explicit --thinking)", async () => {
+	test("supports --model <pattern>:<thinking> (without explicit --thinking)", () => {
 		const registry = {
 			getAll: () => allModels,
 		} as unknown as Parameters<typeof resolveCliModel>[0]["modelRegistry"];
 
-		const result = await resolveCliModel({
+		const result = resolveCliModel({
 			cliModel: "sonnet:high",
 			modelRegistry: registry,
 		});
@@ -253,12 +253,12 @@ describe("resolveCliModel", () => {
 		expect(result.thinkingLevel).toBe("high");
 	});
 
-	test("prefers exact model id match over provider inference (OpenRouter-style ids)", async () => {
+	test("prefers exact model id match over provider inference (OpenRouter-style ids)", () => {
 		const registry = {
 			getAll: () => allModels,
 		} as unknown as Parameters<typeof resolveCliModel>[0]["modelRegistry"];
 
-		const result = await resolveCliModel({
+		const result = resolveCliModel({
 			cliModel: "openai/gpt-4o:extended",
 			modelRegistry: registry,
 		});
@@ -268,12 +268,12 @@ describe("resolveCliModel", () => {
 		expect(result.model?.id).toBe("openai/gpt-4o:extended");
 	});
 
-	test("does not strip invalid :suffix as thinking level in --model (fail fast)", async () => {
+	test("does not strip invalid :suffix as thinking level in --model (fail fast)", () => {
 		const registry = {
 			getAll: () => allModels,
 		} as unknown as Parameters<typeof resolveCliModel>[0]["modelRegistry"];
 
-		const result = await resolveCliModel({
+		const result = resolveCliModel({
 			cliProvider: "openai",
 			cliModel: "gpt-4o:extended",
 			modelRegistry: registry,
@@ -283,12 +283,12 @@ describe("resolveCliModel", () => {
 		expect(result.error).toContain("not found");
 	});
 
-	test("returns a clear error when there are no models", async () => {
+	test("returns a clear error when there are no models", () => {
 		const registry = {
 			getAll: () => [],
 		} as unknown as Parameters<typeof resolveCliModel>[0]["modelRegistry"];
 
-		const result = await resolveCliModel({
+		const result = resolveCliModel({
 			cliProvider: "openai",
 			cliModel: "gpt-4o",
 			modelRegistry: registry,
@@ -298,12 +298,12 @@ describe("resolveCliModel", () => {
 		expect(result.error).toContain("No models available");
 	});
 
-	test("resolves provider-prefixed fuzzy patterns (openrouter/qwen -> openrouter model)", async () => {
+	test("resolves provider-prefixed fuzzy patterns (openrouter/qwen -> openrouter model)", () => {
 		const registry = {
 			getAll: () => allModels,
 		} as unknown as Parameters<typeof resolveCliModel>[0]["modelRegistry"];
 
-		const result = await resolveCliModel({
+		const result = resolveCliModel({
 			cliModel: "openrouter/qwen",
 			modelRegistry: registry,
 		});