@linxiraos/pi-catalog 1.0.0 → 1.0.2

package/CHANGELOG.md

@@ -2,9 +2,81 @@
 
 ## [Unreleased]
 
+## [1.0.2] - 2026-08-15
+
+## [17.3.2] - 2026-08-13
+
+### Added
+
+- Added support for the `deepseek-v4-pro:preview` model
+- Added support for the `gemini-3.7-flash` model
+- Added dynamic Antigravity client-version discovery from the official update manifest (darwin/arm64 channel), so version-gated models appear without a code change; `PI_AI_ANTIGRAVITY_VERSION` remains available as an override.
+
+### Fixed
+
+- Fixed Antigravity discovery missing Gemini 3.7 Flash: Cloud Code Assist gates newer models on the client version in the `User-Agent`, and the pinned `antigravity/hub/2.1.4` was too old. The user-agent now matches the captured 2.8.0 client format (`antigravity/hub/2.8.0 (aidev_client; os_type=darwin; arch=arm64; cl=963137146)`); os_type/arch stay pinned to the darwin/arm64 reference client. Overridable via `PI_AI_ANTIGRAVITY_VERSION` / `PI_AI_ANTIGRAVITY_CL` / `PI_AI_ANTIGRAVITY_OS` / `PI_AI_ANTIGRAVITY_ARCH`.
+
+### Removed
+
+- Removed `ANTIGRAVITY_SYSTEM_INSTRUCTION` from `wire/gemini-headers`; the Antigravity transport and web search no longer inject a fake identity prompt.
+
+## [17.3.1] - 2026-08-13
+
+### Added
+
+- Added dynamic Antigravity and Gemini CLI discovery support for Gemini 3.7 Flash, with low/medium/high thinking-level routing.
+
 ### Changed
 
-- Reset the version to 1.0.0 and republished under the `@linxiraos/*` scope, breaking from the `@oh-my-pi` version lineage.
+- Updated model metadata, context windows, pricing, and configurations in the catalog
+
+## [17.3.0] - 2026-08-13
+
+### Breaking Changes
+
+- Removed `OpenAICompat.enableGeminiThinkingLoopGuard`; thinking-loop eligibility is derived solely from the `model.id` family.
+
+### Added
+
+- Added first-party OpenAI Daybreak Blue, Daybreak Red, and GPT-5.6 Cyber models with full support for their documented API pricing (including long-context rates above 272K input), token limits, tools, and reasoning effort controls (off/low/medium/high/xhigh/max).
+- Added calculateUncachedInputCost() to calculate prompt pricing against active context-length tiers without prompt caching.
+
+### Fixed
+
+- Fixed Anthropic cache-write pricing to correctly honor mixed 5-minute and 1-hour TTL usage instead of incorrectly charging all writes at the 5-minute rate.
+- Fixed Ollama Cloud DeepSeek V4 Flash and older reasoners to correctly apply the DeepSeek effort contract (e.g., low/high/max) instead of the generic effort ladder.
+- Added a default request timeout to OpenAI-compatible model discovery to prevent stalled provider endpoints from hanging startup indefinitely.
+- Fixed Anthropic cache-write pricing to honor mixed 5-minute and 1-hour TTL usage instead of charging every write at the 5-minute rate.
+- Fixed Ollama Cloud DeepSeek V4 Flash (including dated/preview tags like `deepseek-v4-flash:0731`) exposing the generic `minimal`/`low`/`medium`/`high`/`xhigh` effort ladder without `max`; the `ollama-chat` transport now applies the DeepSeek effort contract (Flash → `low`/`high`/`max`, older reasoners → `high`/`max`), matching the direct API and every other host ([#8334](https://github.com/can1357/oh-my-pi/issues/8334)).
+- Exposed the `low` reasoning-effort tier for DeepSeek V4 Pro on the direct API and faithful aggregator routes, matching DeepSeek's updated API contract advertising `reasoning_effort` `low`/`high`/`max` for both V4 SKUs; OpenRouter's non-Flash route still exposes only `high`, and the older V3.x/R1 reasoners remain `high`/`max` ([#8405](https://github.com/can1357/oh-my-pi/issues/8405)).
+- Bounded OpenAI-compatible model discovery with a default request timeout so a stalled provider `/models` endpoint can no longer hang startup indefinitely in `resolveModelDiscoveryFallback` ([#8315](https://github.com/can1357/oh-my-pi/issues/8315)).
+- Fixed Codex-discovered `gpt-daybreak-*` aliases being treated as unknown models, restoring the GPT-5.6 `low`/`medium`/`high`/`xhigh`/`max` effort ladder and its 372K fallback only when the Codex registry omits `context_window`.
+- Fixed first-party OpenAI GPT-5.6 aliases to preserve wire-level `off` through generated pro aliases and to price requests above 272K input at each SKU's documented long-context rates.
+
+## [17.2.15] - 2026-08-12
+
+### Fixed
+
+- Fixed classification of model IDs with large minor versions (e.g., `claude-opus-5-11`) or three-part versions, ensuring they no longer fall back to stale default configurations.
+
+## [17.2.13] - 2026-08-11
+
+### Changed
+
+- Standardized catalog discovery User-Agent headers on `omp/<version>` via the shared `USER_AGENT` utility.
+
+### Fixed
+
+- Marked `meta/muse-spark-1.2` and `muse-spark-1.2-contributor` as image-capable (`input: ["text", "image"]`) with the same Responses reasoning, thinking, and cost metadata as `muse-spark-1.1` (contributor uses its discounted 0.1/0.2 pricing), so `omp models` no longer lists them as text-only.
+- Fixed GLM-5.2 thinking levels across Baseten, CoreWeave, HuggingFace, and other uppercase-ID resellers, which were getting the generic `xhigh` effort ladder instead of the GLM-5.2-specific tiers. Also added Baseten `zai-org/GLM-5.2-Fast` and Fireworks `glm-5.2-fast` as reasoning models ([#8200](https://github.com/can1357/oh-my-pi/pull/8200) by [@jcfrancisco](https://github.com/jcfrancisco)).
+
+## [17.2.12] - 2026-08-08
+
+### Fixed
+
+- Fixed dynamically discovered `alibaba-token-plan/qwen3.8-max` metadata so thinking controls and image input are available ([#8019](https://github.com/can1357/oh-my-pi/issues/8019)).
+- Routed `opencode-go/deepseek-v4-flash` through the OpenAI responses API — the OpenCode Go gateway does not serve this model at `/zen/go/v1/chat/completions`, only at `/zen/go/v1/responses` (`deepseek-v4-pro` keeps chat completions).
+
 ## [17.2.11] - 2026-08-07
 
 ### Fixed
@@ -816,8 +888,8 @@
 
 ### Added
 
-- Added the `ToolCallSyntax` union and `FALLBACK_TOOL_SYNTAX` constant to `@oh-my-pi/pi-catalog/identity` (re-exported from `@oh-my-pi/pi-ai/grammar`).
-- Added `preferredToolSyntax(modelId)` to `@oh-my-pi/pi-catalog/identity`, resolving a model's native tool-call syntax affinity from its family token (Claude→`anthropic`, GLM→`glm`, Kimi→`kimi`, Qwen→`qwen3`, DeepSeek→`deepseek`, OpenAI/gpt-oss→`harmony`, else the `xml` fallback).
+- Added the `ToolCallSyntax` union and `FALLBACK_TOOL_SYNTAX` constant to `@linxiraos/pi-catalog/identity` (re-exported from `@linxiraos/pi-ai/grammar`).
+- Added `preferredToolSyntax(modelId)` to `@linxiraos/pi-catalog/identity`, resolving a model's native tool-call syntax affinity from its family token (Claude→`anthropic`, GLM→`glm`, Kimi→`kimi`, Qwen→`qwen3`, DeepSeek→`deepseek`, OpenAI/gpt-oss→`harmony`, else the `xml` fallback).
 - Added `flux-1-schnell-fp8` to the Fireworks serverless model catalog
 - Added `gpt-oss-20b` to the Fireworks model catalog
 - Added `qwen3-embedding-8b` to the Fireworks model catalog
@@ -835,7 +907,7 @@
 
 ### Added
 
-- Added `modelFamilyToken(modelId)` to `@oh-my-pi/pi-catalog/identity`: a coarse vendor-lineage token (`anthropic`/`openai`/`gemini`/`kimi`/…) for "are two models the same family?" comparisons, backed by `parseKnownModel` canonical-id normalization. Opaque and comparison-only; kind/variant collapsed onto the vendor token ([#2406](https://github.com/can1357/oh-my-pi/issues/2406))
+- Added `modelFamilyToken(modelId)` to `@linxiraos/pi-catalog/identity`: a coarse vendor-lineage token (`anthropic`/`openai`/`gemini`/`kimi`/…) for "are two models the same family?" comparisons, backed by `parseKnownModel` canonical-id normalization. Opaque and comparison-only; kind/variant collapsed onto the vendor token ([#2406](https://github.com/can1357/oh-my-pi/issues/2406))
 
 ### Changed
 
@@ -971,7 +1043,7 @@
 - `buildModel(spec)` (`build.ts`) is now the single Model constructor: it materializes the fully-resolved compat record and canonical thinking metadata exactly once (compat first, thinking derived from identity + resolved compat), so `Model.compat` is a required, complete `CompatOf<TApi>` (`ResolvedOpenAICompat`/`ResolvedOpenAIResponsesCompat`/`ResolvedAnthropicCompat`) and request-path code reads fields with zero URL parsing and zero per-request allocation. Sparse user/config overrides live on the new `ModelSpec<TApi>` input shape and survive on `Model.compatConfig` for introspection.
 - Added `ResolvedAnthropicCompat.supportsSamplingParams` (Opus 4.7+/Fable/Mythos reject `temperature`/`top_p`/`top_k` with a 400), baked at build time from model identity so the request path stops re-parsing model ids.
 - Compat detection gained model-time flags so handlers stop sniffing baseUrl: completions `supportsReasoningParams`, `alwaysSendMaxTokens`, `isOpenRouterHost`, `isVercelGatewayHost`, `streamIdleTimeoutMs`, and a precomputed `whenThinking` alternate view (OpenCode `reasoning_content` gating, #1071/#1484); responses `strictResponsesPairing`, `supportsLongPromptCacheRetention`, `supportsReasoningEffort`; anthropic `officialEndpoint`, `requiresToolResultId`, `replayUnsignedThinking`.
-- New `@oh-my-pi/pi-catalog` package: the model catalog extracted from `@oh-my-pi/pi-ai`. Owns the bundled `models.json` and its generation pipeline (`scripts/generate-models.ts`), the core model data types (`Model`, `Api`, `ThinkingConfig`, `Effort`, `Usage`, compat interfaces), thinking metadata enrichment and generated policies (`model-thinking.ts`), the SQLite model cache and model manager, per-provider discovery factories (`provider-models/`), the discovery protocol clients (`discovery/`), and the new `CATALOG_PROVIDERS` table — the single source of truth for provider ids, default models, and discovery wiring (`KnownProvider`, `PROVIDER_DESCRIPTORS`, and `DEFAULT_MODEL_PER_PROVIDER` are derived from it).
+- New `@linxiraos/pi-catalog` package: the model catalog extracted from `@linxiraos/pi-ai`. Owns the bundled `models.json` and its generation pipeline (`scripts/generate-models.ts`), the core model data types (`Model`, `Api`, `ThinkingConfig`, `Effort`, `Usage`, compat interfaces), thinking metadata enrichment and generated policies (`model-thinking.ts`), the SQLite model cache and model manager, per-provider discovery factories (`provider-models/`), the discovery protocol clients (`discovery/`), and the new `CATALOG_PROVIDERS` table — the single source of truth for provider ids, default models, and discovery wiring (`KnownProvider`, `PROVIDER_DESCRIPTORS`, and `DEFAULT_MODEL_PER_PROVIDER` are derived from it).
 - New `identity/` module centralizing model-identity concerns that were previously duplicated across packages: family classification and version parsing (`identity/classify.ts`, extracted from pi-ai's `model-thinking` internals), canonical model equivalence with injected reference data (`identity/equivalence.ts`, from coding-agent's `model-equivalence`), proxy/reseller reference lookup (`identity/reference.ts`, from coding-agent's `model-registry`), bracket-affix and id-segment helpers (`identity/id.ts`), a single trailing-marker vocabulary with canonical vs reference flavors (`identity/markers.ts` — `search` stays reference-only so Perplexity's `sonar-pro-search` remains canonical-distinct), and provider priority ordering (`identity/priority.ts`).
 - Memoized bundled-reference accessors (`getBundledCanonicalReferenceData` / `getBundledModelReferenceIndex` in `identity/bundled.ts`): one lazy walk of the bundled catalog feeds both canonical equivalence and proxy-reference lookup, so consumers no longer hand-roll the glue.
 - `identity/selection.ts`: pure canonical-variant selection (`resolveCanonicalVariant`, `buildCanonicalModelOrder`, `CanonicalVariantPreferences`) extracted from the coding-agent registry — provider rank, then exact-id match, variant source, id length, and candidate order.
@@ -990,10 +1062,20 @@
 
 - Fixed Anthropic official-endpoint detection to require strict HTTPS hostname matching so non-official or lookalike URLs are no longer treated as official Anthropic hosts
 - Fixed Ollama Cloud dynamic discovery so same-id matches from other providers no longer supply context-window or max-output-token limits for discovered models.
-- Wired `@oh-my-pi/pi-catalog` into the release publish package list, tarball install smoke test, and root `bun generate-models` script.
+- Wired `@linxiraos/pi-catalog` into the release publish package list, tarball install smoke test, and root `bun generate-models` script.
 - Fixed `supportsAdaptiveThinkingDisplay` only matching dash-form version ids: dotted ids (`claude-opus-4.7`) now classify through `identity/classify` like every other anthropic predicate, so six bundled dotted Opus 4.7/4.8 entries (github-copilot, vercel-ai-gateway, zenmux) regain adaptive `display` support; bare dated ids (`claude-opus-4-20250514` = Opus 4.0) stay excluded.
 - Fixed the OpenRouter anthropic adaptive-effort map misclassifying bare dated Opus ids (`claude-opus-4-20250514` parsed as version 4.20 → wrongly adaptive); the map now derives from the shared classifier and the shared 4-/5-tier tables.
 
 ### Removed
 
 - Removed the runtime enrichment layer: `enrichModelThinking` (and its non-enumerable memo-slot cache), `refreshModelThinking`, `modelOmitsReasoningEffort`, and the `model-thinking` re-exports of generator-only policies. Thinking metadata is resolved exactly once inside `buildModel`; runtime helpers (`getSupportedEfforts`, `clampThinkingLevelForModel`, `requireSupportedEffort`, the effort mappers) are pure field reads.
+
+## [1.0.1] - 2026-08-14
+
+- Reset the version to 1.0.0 and republished under the `@linxiraos/*` scope, breaking from the `@oh-my-pi` version lineage.
+
+## [1.0.0] - 2026-08-13
+
+### Changed
+
+- Reset the version to 1.0.0 and republished under the `@linxiraos/*` scope, breaking from the `@oh-my-pi` version lineage.

package/package.json

@@ -1,101 +1,100 @@
 {
-	"type": "module",
-	"name": "@linxiraos/pi-catalog",
-	"version": "1.0.0",
-	"description": "Model catalog for omp: bundled model database, provider discovery descriptors, model identity, classification, and equivalence",
-	"homepage": "https://linxira-os.github.io",
-	"author": "Can Boluk",
-	"license": "MIT",
-	"repository": {
-		"type": "git",
-		"url": "git+https://github.com/Linxira-OS/linxira-zeta.git",
-		"directory": "packages/catalog"
-	},
-	"bugs": {
-		"url": "https://github.com/can1357/oh-my-pi/issues"
-	},
-	"keywords": [
-		"ai",
-		"llm",
-		"models",
-		"catalog",
-		"discovery"
-	],
-	"main": "./src/index.ts",
-	"types": "./dist/types/index.d.ts",
-	"scripts": {
-		"check": "biome check . && bun run check:types",
-		"check:types": "tsgo -p tsconfig.json --noEmit",
-		"lint": "biome lint .",
-		"test": "bun test --parallel",
-		"fix": "biome check --write --unsafe .",
-		"fmt": "biome format --write .",
-		"gen:models": "bun scripts/generate-models.ts",
-		"gen:cursor-proto": "protoc --plugin=protoc-gen-es=../../node_modules/.bin/protoc-gen-es --es_out=src/discovery/cursor-gen --es_opt=target=ts -I ../ai/src/providers/cursor/proto ../ai/src/providers/cursor/proto/agent.proto"
-	},
-	"dependencies": {
-		"@bufbuild/protobuf": "^2.12.1",
-		"@linxiraos/pi-omptype": "1.0.0",
-		"@linxiraos/pi-utils": "1.0.0"
-	},
-	"devDependencies": {
-		"@types/bun": "^1.3.14",
-		"@linxiraos/pi-ai": "1.0.0"
-	},
-	"engines": {
-		"bun": ">=1.3.14"
-	},
-	"files": [
-		"src",
-		"README.md",
-		"CHANGELOG.md",
-		"dist/types"
-	],
-	"exports": {
-		".": {
-			"types": "./dist/types/index.d.ts",
-			"import": "./src/index.ts"
-		},
-		"./models.json": {
-			"types": "./dist/types/models.json.d.d.ts",
-			"import": "./src/models.json"
-		},
-		"./provider-models": {
-			"types": "./dist/types/provider-models/index.d.ts",
-			"import": "./src/provider-models/index.ts"
-		},
-		"./provider-models/*": {
-			"types": "./dist/types/provider-models/*.d.ts",
-			"import": "./src/provider-models/*.ts"
-		},
-		"./discovery": {
-			"types": "./dist/types/discovery/index.d.ts",
-			"import": "./src/discovery/index.ts"
-		},
-		"./discovery/*": {
-			"types": "./dist/types/discovery/*.d.ts",
-			"import": "./src/discovery/*.ts"
-		},
-		"./identity": {
-			"types": "./dist/types/identity/index.d.ts",
-			"import": "./src/identity/index.ts"
-		},
-		"./identity/*": {
-			"types": "./dist/types/identity/*.d.ts",
-			"import": "./src/identity/*.ts"
-		},
-		"./wire/*": {
-			"types": "./dist/types/wire/*.d.ts",
-			"import": "./src/wire/*.ts"
-		},
-		"./compat/*": {
-			"types": "./dist/types/compat/*.d.ts",
-			"import": "./src/compat/*.ts"
-		},
-		"./*": {
-			"types": "./dist/types/*.d.ts",
-			"import": "./src/*.ts"
-		},
-		"./*.js": "./src/*.ts"
-	}
+  "type": "module",
+  "name": "@linxiraos/pi-catalog",
+  "version": "1.0.2",
+  "description": "Model catalog for omp: bundled model database, provider discovery descriptors, model identity, classification, and equivalence",
+  "homepage": "https://linxira-os.github.io",
+  "author": "Can Boluk",
+  "license": "MIT",
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/Linxira-OS/linxira-zeta.git",
+    "directory": "packages/catalog"
+  },
+  "bugs": {
+    "url": "https://github.com/can1357/oh-my-pi/issues"
+  },
+  "keywords": [
+    "ai",
+    "llm",
+    "models",
+    "catalog",
+    "discovery"
+  ],
+  "main": "./src/index.ts",
+  "types": "./src/index.ts",
+  "scripts": {
+    "check": "biome check . && bun run check:types",
+    "check:types": "tsgo -p tsconfig.json --noEmit",
+    "lint": "biome lint .",
+    "test": "bun test --parallel",
+    "fix": "biome check --write --unsafe .",
+    "fmt": "biome format --write .",
+    "gen:models": "bun scripts/generate-models.ts",
+    "gen:cursor-proto": "protoc --plugin=protoc-gen-es=../../node_modules/.bin/protoc-gen-es --es_out=src/discovery/cursor-gen --es_opt=target=ts -I ../ai/src/providers/cursor/proto ../ai/src/providers/cursor/proto/agent.proto"
+  },
+  "dependencies": {
+    "@bufbuild/protobuf": "catalog:",
+    "@linxiraos/pi-omptype": "catalog:",
+    "@linxiraos/pi-utils": "catalog:"
+  },
+  "devDependencies": {
+    "@linxiraos/pi-ai": "catalog:",
+    "@types/bun": "catalog:"
+  },
+  "engines": {
+    "bun": ">=1.3.14"
+  },
+  "files": [
+    "src",
+    "README.md",
+    "CHANGELOG.md"
+  ],
+  "exports": {
+    ".": {
+      "types": "./src/index.ts",
+      "import": "./src/index.ts"
+    },
+    "./models.json": {
+      "types": "./src/models.json.d.ts",
+      "import": "./src/models.json"
+    },
+    "./provider-models": {
+      "types": "./src/provider-models/index.ts",
+      "import": "./src/provider-models/index.ts"
+    },
+    "./provider-models/*": {
+      "types": "./src/provider-models/*.ts",
+      "import": "./src/provider-models/*.ts"
+    },
+    "./discovery": {
+      "types": "./src/discovery/index.ts",
+      "import": "./src/discovery/index.ts"
+    },
+    "./discovery/*": {
+      "types": "./src/discovery/*.ts",
+      "import": "./src/discovery/*.ts"
+    },
+    "./identity": {
+      "types": "./src/identity/index.ts",
+      "import": "./src/identity/index.ts"
+    },
+    "./identity/*": {
+      "types": "./src/identity/*.ts",
+      "import": "./src/identity/*.ts"
+    },
+    "./wire/*": {
+      "types": "./src/wire/*.ts",
+      "import": "./src/wire/*.ts"
+    },
+    "./compat/*": {
+      "types": "./src/compat/*.ts",
+      "import": "./src/compat/*.ts"
+    },
+    "./*": {
+      "types": "./src/*.ts",
+      "import": "./src/*.ts"
+    },
+    "./*.js": "./src/*.ts"
+  }
 }

package/src/build.ts

@@ -14,12 +14,11 @@ import { buildAnthropicCompat } from "./compat/anthropic";
 import { buildBedrockCompat } from "./compat/bedrock";
 import { buildDevinCompat } from "./compat/devin";
 import { buildOpenAICompat, buildOpenAIResponsesCompat, buildOpenRouterCompat } from "./compat/openai";
+import { bareModelId, parseOpenAIModel, semverGte } from "./identity/classify";
 import { resolveModelThinking } from "./model-thinking";
 import type { Api, CompatOf, Model, ModelSpec } from "./types";
 import { cleanModelName } from "./utils";
 
-const OPENAI_GA_COMPUTER_MODEL_RE = /^gpt-5\.(?:[4-9]|[1-9]\d)(?:[.-]|$)/i;
-
 function isDirectOpenAIResponsesEndpoint(spec: ModelSpec<Api>): boolean {
 	if (spec.api === "openai-responses") {
 		if (spec.provider !== "openai") return false;
@@ -55,7 +54,8 @@ function explicitComputerUseConfig(spec: ModelSpec<Api>): boolean | undefined {
 function supportsOpenAIGAComputerUse(spec: ModelSpec<Api>, explicitSupport: boolean | undefined): boolean {
 	if (explicitSupport !== undefined) return explicitSupport;
 	if (!isDirectOpenAIResponsesEndpoint(spec)) return false;
-	return OPENAI_GA_COMPUTER_MODEL_RE.test(spec.requestModelId ?? spec.id);
+	const parsed = parseOpenAIModel(bareModelId(spec.requestModelId ?? spec.id));
+	return parsed !== null && semverGte(parsed.version, "5.4");
 }
 
 export function buildModel<TApi extends Api>(spec: ModelSpec<TApi>): Model<TApi> {

package/src/compat/openai.ts

@@ -22,7 +22,6 @@ import {
 	isMimoModelIdOrName,
 	isOpenAISamplingRestrictedModelId,
 	isQwenModelId,
-	modelFamilyToken,
 } from "../identity/family";
 import type {
 	ModelSpec,
@@ -474,10 +473,6 @@ export function buildOpenAICompat(spec: ModelSpec<"openai-completions">): Resolv
 		supportsSamplingParams: !isOpenAISamplingRestrictedModelId(spec.id),
 		reasoningEffortMap: {},
 		supportsUsageInStreaming: !isCerebras,
-		// pi-ai's thinking-loop guard is gemini-only; default the flag from the
-		// family classifier so OpenAI-compat proxies serving Gemini are covered.
-		// An opaque alias can opt in via `compat.enableGeminiThinkingLoopGuard`.
-		enableGeminiThinkingLoopGuard: modelFamilyToken(spec.id) === "gemini",
 		// Kimi (including via OpenRouter and Fireworks router-form IDs such as
 		// `accounts/fireworks/routers/kimi-*`) calculates TPM rate limits based on
 		// max_tokens, not actual output. The official Kimi K2 model guidance
@@ -749,7 +744,6 @@ export function buildOpenAIResponsesCompat(spec: OpenAIResponsesSpecLike): Resol
 		// lands on Moonshot's MFJS validator.
 		toolSchemaFlavor: isKimiModel ? "moonshot-mfjs" : undefined,
 		alwaysSendMaxTokens: spec.id ? isKimiModelId(spec.id) : false,
-		enableGeminiThinkingLoopGuard: modelFamilyToken(spec.id ?? "") === "gemini",
 		supportsObfuscationOptOut: isOpenAIUrl || spec.provider === "openai",
 		stripDeepseekSpecialTokens:
 			Boolean(id) && isDeepseekModelIdOrName(id) && (spec.provider === "nvidia" || spec.provider === "deepseek"),

package/src/discovery/antigravity.ts

@@ -6,7 +6,7 @@ import {
 	collapseEffortVariants,
 	type VariantCollapseTable,
 } from "../variant-collapse";
-import { getAntigravityUserAgent } from "../wire/gemini-headers";
+import { ensureAntigravityVersion, getAntigravityUserAgent } from "../wire/gemini-headers";
 
 export const ANTIGRAVITY_PRIMARY_ENDPOINT = "https://daily-cloudcode-pa.googleapis.com";
 export const ANTIGRAVITY_SANDBOX_ENDPOINT = "https://daily-cloudcode-pa.sandbox.googleapis.com";
@@ -161,6 +161,10 @@ export interface FetchAntigravityDiscoveryModelsOptions {
 export async function fetchAntigravityDiscoveryModels(
 	options: FetchAntigravityDiscoveryModelsOptions,
 ): Promise<ModelSpec<"google-gemini-cli">[] | null> {
+	if (options.userAgent === undefined) {
+		await ensureAntigravityVersion(options.fetcher ?? fetch, options.signal);
+	}
+
 	const fetcher = discoveryFetch(options.fetcher);
 	const endpoints = options.endpoint
 		? [trimTrailingSlashes(options.endpoint)]

package/src/discovery/openai-compatible.ts

@@ -4,6 +4,18 @@ import { discoveryFetch } from "../utils";
 
 const MODELS_PATH = "/models";
 
+/**
+ * Default hard deadline applied to an OpenAI-compatible `/models` probe when
+ * the caller supplies neither an `AbortSignal` nor an explicit `timeoutMs`.
+ *
+ * Built-in provider model managers (openrouter, xAI, DeepSeek, …) call
+ * {@link fetchOpenAICompatibleModels} with no timeout, so without this bound a
+ * stalled endpoint left the request pending forever and blocked startup's
+ * awaited `resolveModelDiscoveryFallback` discovery pass indefinitely
+ * (issue #8315). 10s matches the coding-agent's remote-discovery budget.
+ */
+export const DEFAULT_OPENAI_COMPATIBLE_DISCOVERY_TIMEOUT_MS = 10_000;
+
 /**
  * Uses a cancellable timer rather than the native abort-timeout helper so
  * successful fast discovery requests do not leave armed timeout signals for
@@ -96,7 +108,11 @@ export interface FetchOpenAICompatibleModelsOptions<TApi extends Api> {
 	headers?: Record<string, string>;
 	/** Optional AbortSignal for request cancellation; caller owns its lifecycle. */
 	signal?: AbortSignal;
-	/** Optional cancellable request timeout used when `signal` is omitted. */
+	/**
+	 * Optional cancellable request timeout used when `signal` is omitted.
+	 * Defaults to {@link DEFAULT_OPENAI_COMPATIBLE_DISCOVERY_TIMEOUT_MS} so a
+	 * stalled endpoint can never hang discovery indefinitely.
+	 */
 	timeoutMs?: number;
 	/** Optional fetch implementation override for testing/custom runtimes. */
 	fetch?: FetchImpl;
@@ -164,9 +180,10 @@ export async function fetchOpenAICompatibleModels<TApi extends Api>(
 	const payload =
 		options.signal !== undefined
 			? await fetchPayload(options.signal)
-			: options.timeoutMs !== undefined
-				? await withOpenAICompatibleDiscoveryTimeout(options.timeoutMs, fetchPayload)
-				: await fetchPayload();
+			: await withOpenAICompatibleDiscoveryTimeout(
+					options.timeoutMs ?? DEFAULT_OPENAI_COMPATIBLE_DISCOVERY_TIMEOUT_MS,
+					fetchPayload,
+				);
 	if (payload === null) {
 		return null;
 	}

package/src/identity/classify.ts

@@ -122,16 +122,31 @@ export const parseAnthropicModel = parser((modelId): AnthropicModel | null => {
 	return { family: "anthropic", kind: kind as AnthropicKind, version };
 });
 
+/**
+ * Rolling OpenAI aliases inherit wire capabilities from their current default
+ * snapshots. Keep this map aligned with the model docs when an alias advances.
+ */
+const OPENAI_ALIAS_VERSIONS: Readonly<Record<string, string>> = {
+	"daybreak-blue-latest": "5.6",
+	"gpt-daybreak-blue-latest": "5.6",
+	"daybreak-red-latest": "5.6",
+	"gpt-daybreak-red-latest": "5.6",
+};
+
 export const parseOpenAIModel = parser((modelId): OpenAIModel | null => {
-	const match = /gpt-(\d+(?:\.\d+){0,2})(?:-(codex-spark|codex-mini|codex-max|codex|mini|max|nano))?\b/.exec(modelId);
-	if (!match) {
+	const aliasVersion = OPENAI_ALIAS_VERSIONS[modelId];
+	const match = aliasVersion
+		? null
+		: /gpt-(\d+(?:\.\d+){0,2})(?:-(codex-spark|codex-mini|codex-max|codex|mini|max|nano))?\b/.exec(modelId);
+	const versionInput = aliasVersion ?? match?.[1];
+	if (!versionInput) {
 		return null;
 	}
-	const version = parseSemVer(match[1]);
+	const version = parseSemVer(versionInput);
 	if (!version) {
 		return null;
 	}
-	return { family: "openai", variant: (match[2] as OpenAIVariant | undefined) ?? "base", version };
+	return { family: "openai", variant: (match?.[2] as OpenAIVariant | undefined) ?? "base", version };
 });
 
 /**
@@ -143,7 +158,7 @@ export const parseOpenAIModel = parser((modelId): OpenAIModel | null => {
  * `parseKnownModel`.
  */
 export const parseGlmModel = parser((modelId): GlmModel | null => {
-	const match = /glm-(\d{1,2}(?:\.\d+)?)(v)?(?:-(air|turbo|flashx|flash|preview))?\b/.exec(modelId);
+	const match = /glm-(\d{1,2}(?:\.\d+)?)(v)?(?:-(air|turbo|flashx|flash|preview))?\b/i.exec(modelId);
 	if (!match) {
 		return null;
 	}
@@ -153,8 +168,8 @@ export const parseGlmModel = parser((modelId): GlmModel | null => {
 	}
 	return {
 		family: "glm",
-		variant: (match[3] as GlmVariant | undefined) ?? "base",
-		vision: match[2] === "v",
+		variant: (match[3]?.toLowerCase() as GlmVariant | undefined) ?? "base",
+		vision: match[2]?.toLowerCase() === "v",
 		version,
 	};
 });
@@ -181,7 +196,9 @@ function createSemVer(major: number, minor: number, patch = 0): SemVer {
 	return { major, minor, patch };
 }
 
-// extend this table if we need anything more than 9.10
+// Fast path for the common 1–2 component versions; anything the table misses
+// (large minors, 3-part versions) parses dynamically below so no future
+// version ever classifies as unknown (the failure class #8256 fixed).
 const precomputeTable: Record<string, SemVer> = {};
 for (let major = 0; major <= 9; major++) {
 	for (let minor = 0; minor <= 10; minor++) {
@@ -192,8 +209,14 @@ for (let major = 0; major <= 9; major++) {
 	precomputeTable[`${major}`] = createSemVer(major, 0, 0);
 }
 
+const SEMVER_PATTERN = /^(\d{1,2})(?:[.-](\d{1,2}))?(?:[.-](\d{1,2}))?$/;
+
 export function parseSemVer(version: string): SemVer | null {
-	return precomputeTable[version] ?? null;
+	const hit = precomputeTable[version];
+	if (hit) return hit;
+	const match = SEMVER_PATTERN.exec(version);
+	if (!match) return null;
+	return createSemVer(Number(match[1]), Number(match[2] ?? 0), Number(match[3] ?? 0));
 }
 
 export function semverGte(left: SemVer | string, right: SemVer | string): boolean {

package/src/identity/family.ts

@@ -85,9 +85,11 @@ export const isDeepseekModelIdOrName = memo((value: string): boolean => {
 
 /**
  * DeepSeek V4 Flash SKU in any host/namespace form (`deepseek-v4-flash`, dated
- * `deepseek-v4-flash-0731`, `deepseek-ai/DeepSeek-V4-Flash`). Flash is the only
- * V4 model whose `reasoning_effort` accepts the `low` tier; V4 Pro tops out at
- * `high`/`max`. See https://api-docs.deepseek.com/api/create-chat-completion.
+ * `deepseek-v4-flash-0731`, `deepseek-ai/DeepSeek-V4-Flash`). Both V4 SKUs
+ * (Flash and Pro) accept the `low` reasoning_effort tier; this predicate keeps
+ * Flash distinguishable from Pro where a host quirk splits them (e.g.
+ * OpenRouter exposes `low` on Flash but only `high` on non-Flash V4).
+ * See https://api-docs.deepseek.com/api/create-chat-completion.
  */
 export const isDeepseekV4FlashModelId = memo((modelId: string): boolean => {
 	return bareModelId(modelId).toLowerCase().includes("deepseek-v4-flash");
@@ -98,6 +100,16 @@ export const isMimoModelIdOrName = memo((value: string): boolean => {
 	return value.toLowerCase().includes("mimo");
 });
 
+/** Gemini family ids in any namespace form (`gemini-*`, `google/gemini-*`, `openrouter/google/gemini-…`). */
+export const isGeminiModelId = memo((modelId: string): boolean => {
+	return /(^|\/)gemini[-.]?/i.test(modelId);
+});
+
+/** Grok family ids across namespace and delimiter forms (`grok-*`, `cursor-grok-*`, `xai/grok-*`). */
+export const isGrokModelId = memo((modelId: string): boolean => {
+	return /(?:^|[./_-])grok(?:[-.]|$)/i.test(modelId);
+});
+
 const GROK_EFFORT_CAPABLE_PREFIXES = ["grok-3-mini", "grok-4.20-multi-agent", "grok-4.3", "grok-4.5"] as const;
 
 /**
@@ -259,12 +271,14 @@ export const modelFamilyToken = memo((modelId: string): string => {
 	const parsed = parseKnownModel(modelId);
 	if (parsed.family !== "unknown") return parsed.family;
 	if (isClaudeModelId(modelId) || isAnthropicNamespacedModelId(modelId)) return "anthropic";
+	if (isGeminiModelId(modelId)) return "gemini";
+	if (isGrokModelId(modelId)) return "grok";
+	if (isDeepseekModelIdOrName(modelId)) return "deepseek";
 	if (isOpenAIModelId(modelId)) return "openai";
 	if (isKimiModelId(modelId)) return "kimi";
 	if (isQwenModelId(modelId)) return "qwen";
 	if (isMinimaxM2FamilyModelId(modelId) || isMinimaxM3FamilyModelId(modelId)) return "minimax";
 	if (isOpenAIGptOssModelId(modelId)) return "gpt-oss";
-	if (isDeepseekModelIdOrName(modelId)) return "deepseek";
 	if (isMimoModelIdOrName(modelId)) return "mimo";
 	if (isGemmaModelId(modelId)) return "gemma";
 	if (parseGlmModel(bareModelId(modelId))) return "glm";

package/src/model-manager.ts

@@ -1,7 +1,7 @@
 import { buildModel } from "./build";
 import { readModelCache, writeModelCache } from "./model-cache";
 import { type GeneratedProvider, getBundledModels } from "./models";
-import type { Api, Model, ModelSpec, Provider } from "./types";
+import type { Api, Model, ModelCost, ModelSpec, Provider, TokenCost } from "./types";
 import { isRecord } from "./utils";
 import { collapseBuiltModelVariants } from "./variant-collapse";
 
@@ -510,6 +510,7 @@ function mergeDynamicModel<TApi extends Api>(existingModel: Model<TApi>, dynamic
 	const reasoning = dynamicReasoningAuthoritative
 		? dynamicModel.reasoning
 		: existingModel.reasoning || dynamicModel.reasoning;
+	const longContextCost = dynamicModel.cost.longContext ?? existingModel.cost.longContext;
 	// Re-build from spec stage: sparse compat comes from `compatConfig` (the
 	// verbatim override vocabulary), never the resolved `compat` record.
 	return buildModel({
@@ -523,6 +524,7 @@ function mergeDynamicModel<TApi extends Api>(existingModel: Model<TApi>, dynamic
 			output: preferDiscoveryCost(dynamicModel.cost.output, existingModel.cost.output),
 			cacheRead: preferDiscoveryCost(dynamicModel.cost.cacheRead, existingModel.cost.cacheRead),
 			cacheWrite: preferDiscoveryCost(dynamicModel.cost.cacheWrite, existingModel.cost.cacheWrite),
+			...(longContextCost ? { longContext: longContextCost } : {}),
 		},
 		contextWindow: preferDiscoveryLimit(dynamicModel.contextWindow, existingModel.contextWindow),
 		maxTokens: preferDiscoveryLimit(dynamicModel.maxTokens, existingModel.maxTokens),
@@ -640,7 +642,7 @@ function isModelInputArray(value: unknown): value is ("text" | "image")[] {
 	return true;
 }
 
-function isModelCost(value: unknown): value is Model<Api>["cost"] {
+function isTokenCost(value: unknown): value is TokenCost {
 	if (!isRecord(value)) {
 		return false;
 	}
@@ -670,3 +672,12 @@ function isModelCost(value: unknown): value is Model<Api>["cost"] {
 	}
 	return true;
 }
+
+function isModelCost(value: unknown): value is ModelCost {
+	if (!isTokenCost(value)) return false;
+	const longContext = (value as TokenCost & { longContext?: unknown }).longContext;
+	if (longContext === undefined) return true;
+	if (!isTokenCost(longContext) || !isRecord(longContext)) return false;
+	const threshold = longContext.inputThreshold;
+	return typeof threshold === "number" && threshold > 0 && threshold < Infinity;
+}