diff --git a/src/assets/__tests__/__snapshots__/assets.snapshot.test.ts.snap b/src/assets/__tests__/__snapshots__/assets.snapshot.test.ts.snap
index a5ba8136..ed954e84 100644
--- a/src/assets/__tests__/__snapshots__/assets.snapshot.test.ts.snap
+++ b/src/assets/__tests__/__snapshots__/assets.snapshot.test.ts.snap
@@ -1357,7 +1357,7 @@ def load_model() -> LLM:
     # CrewAI requires OPENAI_API_KEY env var (ignores api_key parameter)
     os.environ["OPENAI_API_KEY"] = api_key
     return LLM(
-        model="openai/gpt-4o",
+        model="openai/gpt-4.1",
         api_key=api_key
     )
 {{/if}}
@@ -1397,7 +1397,7 @@ def load_model() -> LLM:
     # CrewAI requires GEMINI_API_KEY env var (ignores api_key parameter)
     os.environ["GEMINI_API_KEY"] = api_key
     return LLM(
-        model="gemini/gemini-2.0-flash",
+        model="gemini/gemini-2.5-flash",
         api_key=api_key
     )
 {{/if}}
@@ -1954,7 +1954,7 @@ def _get_api_key() -> str:
 def load_model() -> ChatOpenAI:
     """Get authenticated OpenAI model client."""
     return ChatOpenAI(
-        model="gpt-4o",
+        model="gpt-4.1",
         api_key=_get_api_key()
     )
 {{/if}}
@@ -1991,7 +1991,7 @@ def _get_api_key() -> str:
 def load_model() -> ChatGoogleGenerativeAI:
     """Get authenticated Gemini model client."""
     return ChatGoogleGenerativeAI(
-        model="gemini-2.0-flash",
+        model="gemini-2.5-flash",
         api_key=_get_api_key()
     )
 {{/if}}
@@ -2155,10 +2155,11 @@ async def main(query):
     try:
         async with mcp_server as server:
             active_servers = [server] if server else []
-            # Currently defaults to GPT-4.1
-            # https://openai.github.io/openai-agents-python/models/
             agent = Agent(
-                name="{{ name }}", mcp_servers=active_servers, tools=[add_numbers]
+                name="{{ name }}",
+                model="gpt-4.1",
+                mcp_servers=active_servers,
+                tools=[add_numbers]
             )
             result = await Runner.run(agent, query)
             return result
@@ -2536,7 +2537,7 @@ def load_model() -> OpenAIModel:
     """Get authenticated OpenAI model client."""
     return OpenAIModel(
         client_args={"api_key": _get_api_key()},
-        model_id="gpt-4o",
+        model_id="gpt-4.1",
     )
 {{/if}}
 {{#if (eq modelProvider "Gemini")}}
@@ -2574,7 +2575,7 @@ def load_model() -> GeminiModel:
     """Get authenticated Gemini model client."""
     return GeminiModel(
         client_args={"api_key": _get_api_key()},
-        model_id="gemini-2.0-flash",
+        model_id="gemini-2.5-flash",
     )
 {{/if}}
 "
diff --git a/src/assets/python/crewai/base/model/load.py b/src/assets/python/crewai/base/model/load.py
index 4b08ae53..4e834e42 100644
--- a/src/assets/python/crewai/base/model/load.py
+++ b/src/assets/python/crewai/base/model/load.py
@@ -87,7 +87,7 @@ def load_model() -> LLM:
     # CrewAI requires OPENAI_API_KEY env var (ignores api_key parameter)
     os.environ["OPENAI_API_KEY"] = api_key
     return LLM(
-        model="openai/gpt-4o",
+        model="openai/gpt-4.1",
         api_key=api_key
     )
 {{/if}}
@@ -127,7 +127,7 @@ def load_model() -> LLM:
     # CrewAI requires GEMINI_API_KEY env var (ignores api_key parameter)
     os.environ["GEMINI_API_KEY"] = api_key
     return LLM(
-        model="gemini/gemini-2.0-flash",
+        model="gemini/gemini-2.5-flash",
         api_key=api_key
     )
 {{/if}}
diff --git a/src/assets/python/langchain_langgraph/base/model/load.py b/src/assets/python/langchain_langgraph/base/model/load.py
index 03982c8d..b8f2d71e 100644
--- a/src/assets/python/langchain_langgraph/base/model/load.py
+++ b/src/assets/python/langchain_langgraph/base/model/load.py
@@ -80,7 +80,7 @@ def _get_api_key() -> str:
 def load_model() -> ChatOpenAI:
     """Get authenticated OpenAI model client."""
     return ChatOpenAI(
-        model="gpt-4o",
+        model="gpt-4.1",
         api_key=_get_api_key()
     )
 {{/if}}
@@ -117,7 +117,7 @@ def _get_api_key() -> str:
 def load_model() -> ChatGoogleGenerativeAI:
     """Get authenticated Gemini model client."""
     return ChatGoogleGenerativeAI(
-        model="gemini-2.0-flash",
+        model="gemini-2.5-flash",
         api_key=_get_api_key()
     )
 {{/if}}
diff --git a/src/assets/python/openaiagents/base/main.py b/src/assets/python/openaiagents/base/main.py
index b39c896b..e6b0f0a8 100644
--- a/src/assets/python/openaiagents/base/main.py
+++ b/src/assets/python/openaiagents/base/main.py
@@ -26,10 +26,11 @@ async def main(query):
     try:
         async with mcp_server as server:
             active_servers = [server] if server else []
-            # Currently defaults to GPT-4.1
-            # https://openai.github.io/openai-agents-python/models/
             agent = Agent(
-                name="{{ name }}", mcp_servers=active_servers, tools=[add_numbers]
+                name="{{ name }}",
+                model="gpt-4.1",
+                mcp_servers=active_servers,
+                tools=[add_numbers]
             )
             result = await Runner.run(agent, query)
             return result
diff --git a/src/assets/python/strands/base/model/load.py b/src/assets/python/strands/base/model/load.py
index 95649937..8954269e 100644
--- a/src/assets/python/strands/base/model/load.py
+++ b/src/assets/python/strands/base/model/load.py
@@ -80,7 +80,7 @@ def load_model() -> OpenAIModel:
     """Get authenticated OpenAI model client."""
     return OpenAIModel(
         client_args={"api_key": _get_api_key()},
-        model_id="gpt-4o",
+        model_id="gpt-4.1",
     )
 {{/if}}
 {{#if (eq modelProvider "Gemini")}}
@@ -118,6 +118,6 @@ def load_model() -> GeminiModel:
     """Get authenticated Gemini model client."""
     return GeminiModel(
         client_args={"api_key": _get_api_key()},
-        model_id="gemini-2.0-flash",
+        model_id="gemini-2.5-flash",
     )
 {{/if}}
diff --git a/src/cli/commands/dev/command.tsx b/src/cli/commands/dev/command.tsx
index 63ed6b14..87f5daba 100644
--- a/src/cli/commands/dev/command.tsx
+++ b/src/cli/commands/dev/command.tsx
@@ -58,8 +58,10 @@ export const registerDev = (program: Command) => {
 
           // Determine which agent/port to invoke
           let invokePort = port;
+          let targetAgent = invokeProject?.agents[0];
           if (opts.agent && invokeProject) {
             invokePort = getAgentPort(invokeProject, opts.agent, port);
+            targetAgent = invokeProject.agents.find(a => a.name === opts.agent);
           } else if (invokeProject && invokeProject.agents.length > 1 && !opts.agent) {
             const names = invokeProject.agents.map(a => a.name).join(', ');
             console.error(`Error: Multiple agents found. Use --agent to specify which one.`);
@@ -67,6 +69,11 @@ export const registerDev = (program: Command) => {
             process.exit(1);
           }
 
+          // Show model info if available
+          if (targetAgent?.modelProvider) {
+            console.log(`Provider: ${targetAgent.modelProvider}`);
+          }
+
           await invokeDevServer(invokePort, opts.invoke, opts.stream ?? false);
           return;
         }
@@ -129,8 +136,13 @@ export const registerDev = (program: Command) => {
             console.log(`Port ${basePort} in use, using ${actualPort}`);
           }
 
+          // Get provider info from agent config
+          const targetAgent = project.agents.find(a => a.name === config.agentName);
+          const providerInfo = targetAgent?.modelProvider ?? '(see agent code)';
+
           console.log(`Starting dev server...`);
           console.log(`Agent: ${config.agentName}`);
+          console.log(`Provider: ${providerInfo}`);
           console.log(`Server: http://localhost:${actualPort}/invocations`);
           console.log(`Log: ${logger.getRelativeLogPath()}`);
           console.log(`Press Ctrl+C to stop\n`);
diff --git a/src/cli/commands/invoke/action.ts b/src/cli/commands/invoke/action.ts
index 8d37bafe..19c8e6a5 100644
--- a/src/cli/commands/invoke/action.ts
+++ b/src/cli/commands/invoke/action.ts
@@ -78,6 +78,9 @@ export async function handleInvoke(context: InvokeContext, options: InvokeOption
     return { success: false, error: 'No prompt provided. Usage: agentcore invoke "your prompt"' };
   }
 
+  // Get provider info if available
+  const providerInfo = agentSpec.modelProvider;
+
   // Create logger for this invocation
   const logger = new InvokeLogger({
     agentName: agentSpec.name,
@@ -113,6 +116,7 @@ export async function handleInvoke(context: InvokeContext, options: InvokeOption
         targetName: selectedTargetName,
         response: fullResponse,
         logFilePath: logger.logFilePath,
+        providerInfo,
       };
     } catch (err) {
       logger.logError(err, 'invoke streaming failed');
@@ -136,5 +140,6 @@ export async function handleInvoke(context: InvokeContext, options: InvokeOption
     targetName: selectedTargetName,
     response: response.content,
     logFilePath: logger.logFilePath,
+    providerInfo,
   };
 }
diff --git a/src/cli/commands/invoke/command.tsx b/src/cli/commands/invoke/command.tsx
index 415f124f..36e014c0 100644
--- a/src/cli/commands/invoke/command.tsx
+++ b/src/cli/commands/invoke/command.tsx
@@ -60,7 +60,10 @@ async function handleInvokeCLI(options: InvokeOptions): Promise<void> {
         console.error(`\nLog: ${result.logFilePath}`);
       }
     } else {
-      // Non-streaming, non-json: print response or error
+      // Non-streaming, non-json: print provider info and response or error
+      if (result.providerInfo) {
+        console.error(`Provider: ${result.providerInfo}`);
+      }
       if (result.success && result.response) {
         console.log(result.response);
       } else if (!result.success && result.error) {
diff --git a/src/cli/commands/invoke/types.ts b/src/cli/commands/invoke/types.ts
index b242a83d..c7234bb5 100644
--- a/src/cli/commands/invoke/types.ts
+++ b/src/cli/commands/invoke/types.ts
@@ -14,4 +14,6 @@ export interface InvokeResult {
   response?: string;
   error?: string;
   logFilePath?: string;
+  /** Model provider (e.g., "Anthropic", "Bedrock") */
+  providerInfo?: string;
 }
diff --git a/src/cli/operations/agent/generate/schema-mapper.ts b/src/cli/operations/agent/generate/schema-mapper.ts
index f8fca2e1..ad1a6323 100644
--- a/src/cli/operations/agent/generate/schema-mapper.ts
+++ b/src/cli/operations/agent/generate/schema-mapper.ts
@@ -112,6 +112,7 @@ export function mapGenerateConfigToAgent(config: GenerateConfig): AgentEnvSpec {
     codeLocation: codeLocation as DirectoryPath,
     runtimeVersion: DEFAULT_PYTHON_VERSION,
     networkMode: DEFAULT_NETWORK_MODE,
+    modelProvider: config.modelProvider,
   };
 }
 
diff --git a/src/cli/tui/hooks/useDevServer.ts b/src/cli/tui/hooks/useDevServer.ts
index a2659ba4..569a6a1b 100644
--- a/src/cli/tui/hooks/useDevServer.ts
+++ b/src/cli/tui/hooks/useDevServer.ts
@@ -248,5 +248,6 @@ export function useDevServer(options: { workingDir: string; port: number; agentN
     stop,
     logFilePath: loggerRef.current?.getRelativeLogPath(),
     hasMemory: (project?.memories?.length ?? 0) > 0,
+    modelProvider: project?.agents.find(a => a.name === config?.agentName)?.modelProvider,
   };
 }
diff --git a/src/cli/tui/screens/add/AddFlow.tsx b/src/cli/tui/screens/add/AddFlow.tsx
index 70487f30..236976f3 100644
--- a/src/cli/tui/screens/add/AddFlow.tsx
+++ b/src/cli/tui/screens/add/AddFlow.tsx
@@ -1,3 +1,4 @@
+import { DEFAULT_MODEL_IDS } from '../../../../schema';
 import { computeDefaultCredentialEnvVarName } from '../../../operations/identity/create-identity';
 import { ErrorPrompt } from '../../components';
 import { useAvailableAgents } from '../../hooks/useCreateMcp';
@@ -101,6 +102,11 @@ function AgentAddedSummary({
           </Text>
         )}
       </Box>
+      <Box marginTop={1}>
+        <Text dimColor>Model: </Text>
+        <Text>{DEFAULT_MODEL_IDS[config.modelProvider]}</Text>
+        <Text dimColor> via {config.modelProvider}</Text>
+      </Box>
       {showEnvVarReminder && envVarName && (
         <Box flexDirection="column" marginTop={1}>
           <Text color="yellow">Note: API key not configured.</Text>
diff --git a/src/cli/tui/screens/agent/AddAgentScreen.tsx b/src/cli/tui/screens/agent/AddAgentScreen.tsx
index 0dac24c8..48c5c06e 100644
--- a/src/cli/tui/screens/agent/AddAgentScreen.tsx
+++ b/src/cli/tui/screens/agent/AddAgentScreen.tsx
@@ -1,6 +1,6 @@
 import { APP_DIR, ConfigIO } from '../../../../lib';
 import type { ModelProvider } from '../../../../schema';
-import { AgentNameSchema } from '../../../../schema';
+import { AgentNameSchema, DEFAULT_MODEL_IDS } from '../../../../schema';
 import { computeDefaultCredentialEnvVarName } from '../../../operations/identity/create-identity';
 import {
   ApiKeySecretInput,
@@ -398,7 +398,10 @@ export function AddAgentScreen({ existingAgentNames, onComplete, onExit }: AddAg
               { label: 'Type', value: 'Bring my own code' },
               { label: 'Code Location', value: byoConfig.codeLocation },
               { label: 'Entrypoint', value: byoConfig.entrypoint },
-              { label: 'Model Provider', value: byoConfig.modelProvider },
+              {
+                label: 'Model Provider',
+                value: `${byoConfig.modelProvider} (${DEFAULT_MODEL_IDS[byoConfig.modelProvider]})`,
+              },
               ...(byoConfig.modelProvider !== 'Bedrock'
                 ? [
                     {
diff --git a/src/cli/tui/screens/agent/types.ts b/src/cli/tui/screens/agent/types.ts
index 50e6d05e..bf9b6986 100644
--- a/src/cli/tui/screens/agent/types.ts
+++ b/src/cli/tui/screens/agent/types.ts
@@ -1,5 +1,5 @@
 import type { ModelProvider, PythonRuntime, SDKFramework, TargetLanguage } from '../../../../schema';
-import { getSupportedModelProviders } from '../../../../schema';
+import { DEFAULT_MODEL_IDS, getSupportedModelProviders } from '../../../../schema';
 import type { MemoryOption } from '../generate/types';
 
 // ─────────────────────────────────────────────────────────────────────────────
@@ -89,10 +89,14 @@ export const FRAMEWORK_OPTIONS = [
 ] as const;
 
 export const MODEL_PROVIDER_OPTIONS = [
-  { id: 'Bedrock', title: 'Amazon Bedrock', description: 'AWS managed model inference' },
-  { id: 'Anthropic', title: 'Anthropic', description: 'Claude models via Anthropic API' },
-  { id: 'OpenAI', title: 'OpenAI', description: 'GPT models via OpenAI API' },
-  { id: 'Gemini', title: 'Google Gemini', description: 'Gemini models via Google API' },
+  { id: 'Bedrock', title: `Amazon Bedrock (${DEFAULT_MODEL_IDS.Bedrock})`, description: 'AWS managed model inference' },
+  {
+    id: 'Anthropic',
+    title: `Anthropic (${DEFAULT_MODEL_IDS.Anthropic})`,
+    description: 'Claude models via Anthropic API',
+  },
+  { id: 'OpenAI', title: `OpenAI (${DEFAULT_MODEL_IDS.OpenAI})`, description: 'GPT models via OpenAI API' },
+  { id: 'Gemini', title: `Google Gemini (${DEFAULT_MODEL_IDS.Gemini})`, description: 'Gemini models via Google API' },
 ] as const;
 
 /**
diff --git a/src/cli/tui/screens/create/CreateScreen.tsx b/src/cli/tui/screens/create/CreateScreen.tsx
index 84ae7f55..79a62c84 100644
--- a/src/cli/tui/screens/create/CreateScreen.tsx
+++ b/src/cli/tui/screens/create/CreateScreen.tsx
@@ -1,4 +1,4 @@
-import { ProjectNameSchema } from '../../../../schema';
+import { DEFAULT_MODEL_IDS, ProjectNameSchema } from '../../../../schema';
 import { validateFolderNotExists } from '../../../commands/create/validate';
 import { computeDefaultCredentialEnvVarName } from '../../../operations/identity/create-identity';
 import {
@@ -48,11 +48,14 @@ function buildExitMessage(projectName: string, steps: Step[], agentConfig: AddAg
   if (agentConfig?.agentType === 'create') {
     const frameworkOption = FRAMEWORK_OPTIONS.find(o => o.id === agentConfig.framework);
     const frameworkLabel = frameworkOption?.title ?? agentConfig.framework;
+    const modelName = DEFAULT_MODEL_IDS[agentConfig.modelProvider];
     const agentPath = `app/${agentConfig.name}/`;
     const agentcorePath = 'agentcore/';
     const maxPathLen = Math.max(agentPath.length, agentcorePath.length);
     lines.push(`    ${agentPath.padEnd(maxPathLen)}  \x1b[2m${agentConfig.language} agent (${frameworkLabel})\x1b[0m`);
     lines.push(`    ${agentcorePath.padEnd(maxPathLen)}  \x1b[2mConfig and CDK project\x1b[0m`);
+    lines.push('');
+    lines.push(`\x1b[2mModel:\x1b[0m ${modelName} \x1b[2mvia ${agentConfig.modelProvider}\x1b[0m`);
   } else if (agentConfig?.agentType === 'byo') {
     const agentPath = agentConfig.codeLocation;
     const agentcorePath = 'agentcore/';
@@ -169,6 +172,13 @@ function CreatedSummary({ projectName, agentConfig }: { projectName: string; age
           </Text>
         </Box>
       </Box>
+      {isCreate && agentConfig && (
+        <Box marginTop={1}>
+          <Text dimColor>Model: </Text>
+          <Text>{DEFAULT_MODEL_IDS[agentConfig.modelProvider]}</Text>
+          <Text dimColor> via {agentConfig.modelProvider}</Text>
+        </Box>
+      )}
       {isByo && agentConfig && (
         <Box flexDirection="column" marginTop={1}>
           <Text color="yellow">
diff --git a/src/cli/tui/screens/dev/DevScreen.tsx b/src/cli/tui/screens/dev/DevScreen.tsx
index e376551f..b4d91338 100644
--- a/src/cli/tui/screens/dev/DevScreen.tsx
+++ b/src/cli/tui/screens/dev/DevScreen.tsx
@@ -160,6 +160,7 @@ export function DevScreen(props: DevScreenProps) {
     stop,
     logFilePath,
     hasMemory,
+    modelProvider,
   } = useDevServer({
     workingDir,
     port: props.port ?? 8080,
@@ -383,6 +384,12 @@ export function DevScreen(props: DevScreenProps) {
         <Text>Agent: </Text>
         <Text color="green">{config?.agentName}</Text>
       </Box>
+      {modelProvider && (
+        <Box>
+          <Text>Provider: </Text>
+          <Text color="green">{modelProvider}</Text>
+        </Box>
+      )}
       <Box>
         <Text>Server: </Text>
         <Text color="cyan">http://localhost:{actualPort}/invocations</Text>
diff --git a/src/cli/tui/screens/generate/GenerateWizardUI.tsx b/src/cli/tui/screens/generate/GenerateWizardUI.tsx
index 9947cf05..a80753f5 100644
--- a/src/cli/tui/screens/generate/GenerateWizardUI.tsx
+++ b/src/cli/tui/screens/generate/GenerateWizardUI.tsx
@@ -1,5 +1,5 @@
 import type { ModelProvider } from '../../../../schema';
-import { ProjectNameSchema } from '../../../../schema';
+import { DEFAULT_MODEL_IDS, ProjectNameSchema } from '../../../../schema';
 import { computeDefaultCredentialEnvVarName } from '../../../operations/identity/create-identity';
 import { ApiKeySecretInput, Panel, SelectList, StepIndicator, TextInput } from '../../components';
 import type { SelectableItem } from '../../components';
@@ -202,7 +202,9 @@ function ConfirmView({ config, credentialProjectName }: { config: GenerateConfig
         </Text>
         <Text>
           <Text dimColor>Model Provider: </Text>
-          <Text>{config.modelProvider}</Text>
+          <Text>
+            {config.modelProvider} ({DEFAULT_MODEL_IDS[config.modelProvider]})
+          </Text>
         </Text>
         {config.modelProvider !== 'Bedrock' && (
           <Text>
diff --git a/src/cli/tui/screens/generate/types.ts b/src/cli/tui/screens/generate/types.ts
index f27f4574..5c1510ed 100644
--- a/src/cli/tui/screens/generate/types.ts
+++ b/src/cli/tui/screens/generate/types.ts
@@ -1,5 +1,5 @@
 import type { ModelProvider, SDKFramework, TargetLanguage } from '../../../../schema';
-import { getSupportedModelProviders } from '../../../../schema';
+import { DEFAULT_MODEL_IDS, getSupportedModelProviders } from '../../../../schema';
 
 export type GenerateStep = 'projectName' | 'language' | 'sdk' | 'modelProvider' | 'apiKey' | 'memory' | 'confirm';
 
@@ -51,10 +51,14 @@ export const SDK_OPTIONS = [
 ] as const;
 
 export const MODEL_PROVIDER_OPTIONS = [
-  { id: 'Bedrock', title: 'Amazon Bedrock', description: 'AWS managed model inference' },
-  { id: 'Anthropic', title: 'Anthropic', description: 'Claude models via Anthropic API' },
-  { id: 'OpenAI', title: 'OpenAI', description: 'GPT models via OpenAI API' },
-  { id: 'Gemini', title: 'Google Gemini', description: 'Gemini models via Google API' },
+  { id: 'Bedrock', title: `Amazon Bedrock (${DEFAULT_MODEL_IDS.Bedrock})`, description: 'AWS managed model inference' },
+  {
+    id: 'Anthropic',
+    title: `Anthropic (${DEFAULT_MODEL_IDS.Anthropic})`,
+    description: 'Claude models via Anthropic API',
+  },
+  { id: 'OpenAI', title: `OpenAI (${DEFAULT_MODEL_IDS.OpenAI})`, description: 'GPT models via OpenAI API' },
+  { id: 'Gemini', title: `Google Gemini (${DEFAULT_MODEL_IDS.Gemini})`, description: 'Gemini models via Google API' },
 ] as const;
 
 /**
diff --git a/src/cli/tui/screens/invoke/InvokeScreen.tsx b/src/cli/tui/screens/invoke/InvokeScreen.tsx
index ed9a2319..ad481605 100644
--- a/src/cli/tui/screens/invoke/InvokeScreen.tsx
+++ b/src/cli/tui/screens/invoke/InvokeScreen.tsx
@@ -285,6 +285,12 @@ export function InvokeScreen({
           <Text color="cyan">{agent?.name}</Text>
         </Box>
       )}
+      {mode !== 'select-agent' && agent?.modelProvider && (
+        <Box>
+          <Text>Provider: </Text>
+          <Text color="cyan">{agent.modelProvider}</Text>
+        </Box>
+      )}
       <Box>
         <Text>Target: </Text>
         <Text color="yellow">{config.target.region}</Text>
diff --git a/src/cli/tui/screens/invoke/useInvokeFlow.ts b/src/cli/tui/screens/invoke/useInvokeFlow.ts
index 131122b7..74c49ed7 100644
--- a/src/cli/tui/screens/invoke/useInvokeFlow.ts
+++ b/src/cli/tui/screens/invoke/useInvokeFlow.ts
@@ -2,6 +2,7 @@ import { ConfigIO } from '../../../../lib';
 import type {
   AgentCoreDeployedState,
   AwsDeploymentTarget,
+  ModelProvider,
   AgentCoreProjectSpec as _AgentCoreProjectSpec,
 } from '../../../../schema';
 import { invokeAgentRuntimeStreaming } from '../../../aws';
@@ -11,7 +12,7 @@ import { generateSessionId } from '../../../operations/session';
 import { useCallback, useEffect, useRef, useState } from 'react';
 
 export interface InvokeConfig {
-  agents: { name: string; state: AgentCoreDeployedState }[];
+  agents: { name: string; state: AgentCoreDeployedState; modelProvider?: ModelProvider }[];
   target: AwsDeploymentTarget;
   targetName: string;
   projectName: string;
@@ -77,7 +78,7 @@ export function useInvokeFlow(options: InvokeFlowOptions = {}): InvokeFlowState
         for (const agent of project.agents) {
           const state = targetState?.resources?.agents?.[agent.name];
           if (state) {
-            agents.push({ name: agent.name, state });
+            agents.push({ name: agent.name, state, modelProvider: agent.modelProvider });
           }
         }
 
diff --git a/src/schema/constants.ts b/src/schema/constants.ts
index 7ab4bb60..cff05c1b 100644
--- a/src/schema/constants.ts
+++ b/src/schema/constants.ts
@@ -13,6 +13,17 @@ export type TargetLanguage = z.infer<typeof TargetLanguageSchema>;
 export const ModelProviderSchema = z.enum(['Bedrock', 'Gemini', 'OpenAI', 'Anthropic']);
 export type ModelProvider = z.infer<typeof ModelProviderSchema>;
 
+/**
+ * Default model IDs used for each provider.
+ * These are the models generated in agent templates.
+ */
+export const DEFAULT_MODEL_IDS: Record<ModelProvider, string> = {
+  Bedrock: 'us.anthropic.claude-sonnet-4-5-20250514-v1:0',
+  Anthropic: 'claude-sonnet-4-5-20250514',
+  OpenAI: 'gpt-4.1',
+  Gemini: 'gemini-2.5-flash',
+};
+
 /**
  * Matrix defining which model providers are supported for each SDK framework.
  * - Most SDKs support all 4 providers (Bedrock, Anthropic, OpenAI, Gemini)
diff --git a/src/schema/schemas/agent-env.ts b/src/schema/schemas/agent-env.ts
index f0c72d9d..9f01a8e0 100644
--- a/src/schema/schemas/agent-env.ts
+++ b/src/schema/schemas/agent-env.ts
@@ -3,7 +3,11 @@
  *
  * @module agent-env
  */
-import { NetworkModeSchema, RuntimeVersionSchema as RuntimeVersionSchemaFromConstants } from '../constants';
+import {
+  ModelProviderSchema,
+  NetworkModeSchema,
+  RuntimeVersionSchema as RuntimeVersionSchemaFromConstants,
+} from '../constants';
 import type { DirectoryPath, FilePath } from '../types';
 import { z } from 'zod';
 
@@ -111,6 +115,8 @@ export const AgentEnvSpecSchema = z.object({
   networkMode: NetworkModeSchema.optional(),
   /** Instrumentation settings for observability. Defaults to OTel enabled. */
   instrumentation: InstrumentationSchema.optional(),
+  /** Model provider used by this agent. Optional for backwards compatibility. */
+  modelProvider: ModelProviderSchema.optional(),
 });
 
 export type AgentEnvSpec = z.infer<typeof AgentEnvSpecSchema>;