feat: use provider defaults for memory models (#27129)

## Why

Memory startup used hardcoded OpenAI model slugs for extraction and
consolidation. That works for the default OpenAI-compatible path, but
provider-specific backends can require different model identifiers. In
particular, Amazon Bedrock should use its Bedrock model ID for these
background memory requests instead of the OpenAI `gpt-5.4-mini` /
`gpt-5.4` slugs.

## What Changed

- Added provider-owned preferred memory model methods alongside
`approval_review_preferred_model`.
- Updated memory extraction and consolidation to resolve their default
model through the active `ModelProvider`.
- Added Amazon Bedrock overrides so both memory stages use
`openai.gpt-5.4` through Bedrock’s provider-specific model ID.
- Kept explicit `memories.extract_model` and
`memories.consolidation_model` config overrides taking precedence.
- Added startup coverage for default OpenAI and Bedrock memory model
selection.

#closes #26288

codex-rs/model-provider/src/amazon_bedrock/mod.rs

@@ -68,6 +68,14 @@ impl ModelProvider for AmazonBedrockModelProvider {
         AMAZON_BEDROCK_GPT_5_4_MODEL_ID
     }
 
+    fn memory_extraction_preferred_model(&self) -> &'static str {
+        AMAZON_BEDROCK_GPT_5_4_MODEL_ID
+    }
+
+    fn memory_consolidation_preferred_model(&self) -> &'static str {
+        AMAZON_BEDROCK_GPT_5_4_MODEL_ID
+    }
+
     fn auth_manager(&self) -> Option<Arc<AuthManager>> {
         None
     }

codex-rs/model-provider/src/provider.rs

@@ -74,6 +74,14 @@ pub type ProviderAccountResult = std::result::Result<ProviderAccountState, Provi
 /// require a backend-specific model ID.
 pub const DEFAULT_APPROVAL_REVIEW_PREFERRED_MODEL: &str = "codex-auto-review";
 
+/// Default model used for memory extraction when a provider does not require a
+/// backend-specific model ID.
+pub const DEFAULT_MEMORY_EXTRACTION_PREFERRED_MODEL: &str = "gpt-5.4-mini";
+
+/// Default model used for memory consolidation when a provider does not require
+/// a backend-specific model ID.
+pub const DEFAULT_MEMORY_CONSOLIDATION_PREFERRED_MODEL: &str = "gpt-5.4";
+
 /// Runtime provider abstraction used by model execution.
 ///
 /// Implementations own provider-specific behavior for a model backend. The
@@ -96,6 +104,20 @@ pub trait ModelProvider: fmt::Debug + Send + Sync {
         DEFAULT_APPROVAL_REVIEW_PREFERRED_MODEL
     }
 
+    /// Returns the preferred model used for memory extraction.
+    ///
+    /// Providers that require backend-specific model IDs should override this.
+    fn memory_extraction_preferred_model(&self) -> &'static str {
+        DEFAULT_MEMORY_EXTRACTION_PREFERRED_MODEL
+    }
+
+    /// Returns the preferred model used for memory consolidation.
+    ///
+    /// Providers that require backend-specific model IDs should override this.
+    fn memory_consolidation_preferred_model(&self) -> &'static str {
+        DEFAULT_MEMORY_CONSOLIDATION_PREFERRED_MODEL
+    }
+
     /// Returns whether requests made through this provider should include attestation.
     fn supports_attestation(&self) -> bool {
         false