Expose multi-agent versions in model/list (#37433)

## What changed

- Add nullable `multiAgentVersion` metadata to v2 `model/list` responses.
- Represent supported values as `disabled`, `v1`, or `v2` in the Rust API and generated JSON Schema and TypeScript exports.
- Forward the value from model catalog presets through app-server responses.

## Testing

- Extend the remote model catalog test fixture with a `v2` multi-agent version.

GitOrigin-RevId: bd3b2bbe0a3b562b4d1f03df7da2a73a35713978

codex-rs/app-server/README.md

@@ -228,7 +228,7 @@ Example with notification opt-out:
 - `fs/watch` — subscribe this connection to filesystem change notifications for an absolute file or directory path and caller-provided `watchId`; returns the canonicalized `path`.
 - `fs/unwatch` — stop sending notifications for a prior `fs/watch`; returns `{}`.
 - `fs/changed` — notification emitted when watched paths change, including the `watchId` and `changedPaths`.
-- `model/list` — list available models (set `includeHidden: true` to include entries with `hidden: true`), with model-advertised string reasoning effort options in the catalog's intended progression order, optional `modelSpecialty`, `additionalSpeedTiers`, `serviceTiers`, optional `defaultServiceTier`, optional legacy `upgrade` model ids, optional `upgradeInfo` metadata (`model`, `upgradeCopy`, `modelLink`, `migrationMarkdown`), and optional `availabilityNux` metadata. Clients should preserve the `supportedReasoningEfforts` array order rather than deriving order from the effort names.
+- `model/list` — list available models (set `includeHidden: true` to include entries with `hidden: true`), with model-advertised string reasoning effort options in the catalog's intended progression order, optional `modelSpecialty`, nullable `multiAgentVersion` (`disabled`, `v1`, or `v2`), `additionalSpeedTiers`, `serviceTiers`, optional `defaultServiceTier`, optional legacy `upgrade` model ids, optional `upgradeInfo` metadata (`model`, `upgradeCopy`, `modelLink`, `migrationMarkdown`), and optional `availabilityNux` metadata. Clients should preserve the `supportedReasoningEfforts` array order rather than deriving order from the effort names.
 - `modelProvider/capabilities/read` — read provider-level capabilities for the currently configured model provider.
 - `experimentalFeature/list` — list feature flags with stage metadata (`beta`, `underDevelopment`, `stable`, etc.), enabled/default-enabled state, and cursor pagination. Pass `threadId` when showing feature state for an existing loaded thread so `enabled` is computed from that thread's refreshed config, including project-local config for the thread's cwd; if omitted, the server uses its default config resolution context. For non-beta flags, `displayName`/`description`/`announcement` are `null`.
 - `permissionProfile/list` — beta; list available permission profile ids with optional display `description` text and an `allowed` flag reflecting effective requirements, using cursor pagination. Pass `cwd` when the caller needs project-local `[permissions.<id>]` entries to be included in the current catalog view.

codex-rs/app-server/src/models.rs

@@ -46,6 +46,7 @@ fn model_from_preset(preset: ModelPreset) -> Model {
         default_reasoning_effort: preset.default_reasoning_effort,
         input_modalities: preset.input_modalities,
         supports_personality: preset.supports_personality,
+        multi_agent_version: preset.multi_agent_version.map(Into::into),
         additional_speed_tiers: preset.additional_speed_tiers,
         service_tiers: preset
             .service_tiers

codex-rs/app-server/tests/common/models_cache.rs

@@ -68,7 +68,7 @@ fn preset_to_info(preset: &ModelPreset, priority: i32) -> ModelInfo {
         auto_review_model_override: None,
         model_specialty: None,
         tool_mode: None,
-        multi_agent_version: None,
+        multi_agent_version: preset.multi_agent_version,
     }
 }
 

codex-rs/app-server/tests/suite/v2/model_list.rs

@@ -61,6 +61,7 @@ fn model_from_preset(preset: &ModelPreset) -> Model {
         // cache report `supports_personality = false`.
         // todo(sayan): fix, maybe make roundtrip use ModelInfo only
         supports_personality: false,
+        multi_agent_version: preset.multi_agent_version.map(Into::into),
         additional_speed_tiers: preset.additional_speed_tiers.clone(),
         service_tiers: preset
             .service_tiers
@@ -177,6 +178,7 @@ async fn list_models_uses_chatgpt_remote_catalog_as_source_of_truth() -> Result<
         "truncation_policy": {"mode": "bytes", "limit": 10_000},
         "supports_parallel_tool_calls": false,
         "supports_image_detail_original": false,
+        "multi_agent_version": "v2",
         "context_window": 272_000,
         "max_context_window": 272_000,
         "experimental_supported_tools": [],