@molecule/api-resource-ai-models 1.11.0 → 1.12.0

package/README.md

@@ -3,7 +3,7 @@ AUTO-GENERATED — DO NOT EDIT THIS FILE.
 Generated by `mlcl sync-docs` from the package's src/index.ts JSDoc + mlcl/registry.json.
 Edits here are overwritten on the next commit (molecule's pre-commit hook regenerates).
 To change this document, edit the module-level JSDoc in src/index.ts.
-Generated: 2026-09-30T02:00:14.123Z
+Generated: 2026-09-30T04:56:41.977Z
 -->
 
 # @molecule/api-resource-ai-models
@@ -966,11 +966,20 @@ Sources (verified 2026-07-28; OpenAI re-verified 2026-07-31 after the
   low|medium|high, vision, tools, caching, search grounding, code execution,
   url context — identical surface to 3.7-flash. Knowledge cutoff is published
   by neither Google nor models.dev for this id)
-- xAI: https://docs.x.ai/developers/models + /developers/grok-4-5
-  (grok-4.5 flagship 2026-07-08: $2/$6, 500K ctx, ≥200K prompts bill 2× —
-  not modeled; reasoning_effort low|medium|high default high, image input;
-  grok-4.3 still served at $1.25/$2.50 with the bigger 1M window;
-  grok-code-fast-1 no longer listed — retires 2026-08-15)
+- xAI: https://docs.x.ai/developers/models + /developers/grok-4-7 +
+  /developers/model-capabilities/text/reasoning (re-read 2026-09-30:
+  grok-4.7 flagship: $2/$6, cached $0.50, 500K ctx, ≥200K prompts bill 2× —
+  not modeled; reasoning_effort low|medium|high|xhigh default high, image
+  input, knowledge cutoff May 2026, "No text output limit"; live-probed on
+  /v1/chat/completions with tools, forced + required tool_choice,
+  temperature 0, xhigh and an image — all 200, effort 'none' 400s.
+  grok-4.6 (re-read 2026-09-30 from the models page's model list — its own
+  page now redirects to grok-4-7; specs from that page as archived
+  2026-09-10): $2/$6, cached $0.50, 500K ctx, image input, reasoning_effort
+  low|medium|high|xhigh default high, knowledge cutoff February 1, 2026 —
+  cataloged superseded by grok-4.7.
+  grok-4.5 still served at $2/$6, cached $0.30; grok-4.3 still served at
+  $1.25/$2.50 with the bigger 1M window; grok-code-fast-1 no longer listed)
 - Chinese public holidays (DeepSeek's peak excludes them):
   https://www.12371.gov.cn/web/article/web/content_1451614684968525824.html
   (the State Council's 2026 notice; see CHINA_PUBLIC_HOLIDAYS)

package/dist/models.d.ts

@@ -182,11 +182,20 @@ import type { ModelDefinition } from './types.js';
  *   low|medium|high, vision, tools, caching, search grounding, code execution,
  *   url context — identical surface to 3.7-flash. Knowledge cutoff is published
  *   by neither Google nor models.dev for this id)
- * - xAI: https://docs.x.ai/developers/models + /developers/grok-4-5
- *   (grok-4.5 flagship 2026-07-08: $2/$6, 500K ctx, ≥200K prompts bill 2× —
- *   not modeled; reasoning_effort low|medium|high default high, image input;
- *   grok-4.3 still served at $1.25/$2.50 with the bigger 1M window;
- *   grok-code-fast-1 no longer listed — retires 2026-08-15)
+ * - xAI: https://docs.x.ai/developers/models + /developers/grok-4-7 +
+ *   /developers/model-capabilities/text/reasoning (re-read 2026-09-30:
+ *   grok-4.7 flagship: $2/$6, cached $0.50, 500K ctx, ≥200K prompts bill 2× —
+ *   not modeled; reasoning_effort low|medium|high|xhigh default high, image
+ *   input, knowledge cutoff May 2026, "No text output limit"; live-probed on
+ *   /v1/chat/completions with tools, forced + required tool_choice,
+ *   temperature 0, xhigh and an image — all 200, effort 'none' 400s.
+ *   grok-4.6 (re-read 2026-09-30 from the models page's model list — its own
+ *   page now redirects to grok-4-7; specs from that page as archived
+ *   2026-09-10): $2/$6, cached $0.50, 500K ctx, image input, reasoning_effort
+ *   low|medium|high|xhigh default high, knowledge cutoff February 1, 2026 —
+ *   cataloged superseded by grok-4.7.
+ *   grok-4.5 still served at $2/$6, cached $0.30; grok-4.3 still served at
+ *   $1.25/$2.50 with the bigger 1M window; grok-code-fast-1 no longer listed)
  * - Chinese public holidays (DeepSeek's peak excludes them):
  *   https://www.12371.gov.cn/web/article/web/content_1451614684968525824.html
  *   (the State Council's 2026 notice; see CHINA_PUBLIC_HOLIDAYS)

package/dist/models.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"models.d.ts","sourceRoot":"","sources":["../src/models.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,YAAY,CAAA;AAwDjD;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA6RG;AACH,eAAO,MAAM,MAAM,EAAE,SAAS,eAAe,EA2sEnC,CAAA"}
+{"version":3,"file":"models.d.ts","sourceRoot":"","sources":["../src/models.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,YAAY,CAAA;AAwDjD;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAsSG;AACH,eAAO,MAAM,MAAM,EAAE,SAAS,eAAe,EAgxEnC,CAAA"}

package/dist/models.js

@@ -231,11 +231,20 @@ const CHINA_PUBLIC_HOLIDAYS = [
  *   low|medium|high, vision, tools, caching, search grounding, code execution,
  *   url context — identical surface to 3.7-flash. Knowledge cutoff is published
  *   by neither Google nor models.dev for this id)
- * - xAI: https://docs.x.ai/developers/models + /developers/grok-4-5
- *   (grok-4.5 flagship 2026-07-08: $2/$6, 500K ctx, ≥200K prompts bill 2× —
- *   not modeled; reasoning_effort low|medium|high default high, image input;
- *   grok-4.3 still served at $1.25/$2.50 with the bigger 1M window;
- *   grok-code-fast-1 no longer listed — retires 2026-08-15)
+ * - xAI: https://docs.x.ai/developers/models + /developers/grok-4-7 +
+ *   /developers/model-capabilities/text/reasoning (re-read 2026-09-30:
+ *   grok-4.7 flagship: $2/$6, cached $0.50, 500K ctx, ≥200K prompts bill 2× —
+ *   not modeled; reasoning_effort low|medium|high|xhigh default high, image
+ *   input, knowledge cutoff May 2026, "No text output limit"; live-probed on
+ *   /v1/chat/completions with tools, forced + required tool_choice,
+ *   temperature 0, xhigh and an image — all 200, effort 'none' 400s.
+ *   grok-4.6 (re-read 2026-09-30 from the models page's model list — its own
+ *   page now redirects to grok-4-7; specs from that page as archived
+ *   2026-09-10): $2/$6, cached $0.50, 500K ctx, image input, reasoning_effort
+ *   low|medium|high|xhigh default high, knowledge cutoff February 1, 2026 —
+ *   cataloged superseded by grok-4.7.
+ *   grok-4.5 still served at $2/$6, cached $0.30; grok-4.3 still served at
+ *   $1.25/$2.50 with the bigger 1M window; grok-code-fast-1 no longer listed)
  * - Chinese public holidays (DeepSeek's peak excludes them):
  *   https://www.12371.gov.cn/web/article/web/content_1451614684968525824.html
  *   (the State Council's 2026 notice; see CHINA_PUBLIC_HOLIDAYS)
@@ -1435,13 +1444,78 @@ export const MODELS = [
     },
     // ---------------------------------------------------------------------------
     // xAI (Grok)
-    // Verified: https://docs.x.ai/developers/models + /developers/grok-4-5
-    //           (2026-07-28)
-    // grok-4.5 (2026-07-08) is the flagship: 500K ctx, reasoning_effort
-    // low|medium|high default high, image input. grok-4.3 stays served as the
-    // value tier with the BIGGER 1M window (reasoning_effort none|low|medium|
-    // high, default low). Max output tokens still not documented for any model.
+    // Verified: https://docs.x.ai/developers/models + /developers/grok-4-7 +
+    //           /developers/model-capabilities/text/reasoning (2026-09-30)
+    // grok-4.7 is the flagship: 500K ctx, reasoning_effort low|medium|high|xhigh
+    // default high ("xhigh is available on grok-4.6 and later"), image input,
+    // knowledge cutoff May 2026. grok-4.5 is still served but superseded.
+    // grok-4.3 stays served with the BIGGER 1M window (reasoning_effort
+    // none|low|medium|high, default low). Max output tokens still not documented
+    // for any model (grok-4.7's page says "No text output limit").
     // ---------------------------------------------------------------------------
+    {
+        id: 'grok-4.7',
+        provider: 'xai',
+        label: 'Grok 4.7',
+        description: 'xAI frontier — coding, agentic tasks & knowledge work',
+        contextWindow: 500_000,
+        // "No text output limit" per docs.x.ai — conservative cap.
+        maxOutputTokens: 128_000,
+        supportsThinking: true,
+        thinkingBudgetTokens: 16_000,
+        thinkingConfigurable: true,
+        // reasoning_effort low|medium|high|xhigh, default high (docs.x.ai/developers/
+        // grok-4-7). No 'none' tier — the live endpoint 400s on it (2026-09-30).
+        supportedEffortLevels: ['low', 'medium', 'high', 'xhigh'],
+        defaultEffortLevel: 'high',
+        // Multimodal input (text + images), text out.
+        supportsVision: true,
+        supportsPromptCaching: true,
+        supportsTools: true,
+        // <200K-token prompts; ≥200K bills 2× ($4/$1.00/$12 — tiering not
+        // modeled, same as grok-4.5 and the Gemini 3.1 Pro >200K tier).
+        inputPricePerMTok: 2,
+        outputPricePerMTok: 6,
+        // xAI cached input billed at a flat $0.50/M (up from 4.5's $0.30), no
+        // write premium.
+        cacheReadPricePerMTok: 0.5,
+        cacheWritePricePerMTok: 2,
+        // Official (docs.x.ai): May 2026.
+        knowledgeCutoff: '2026-05-01',
+    },
+    {
+        id: 'grok-4.6',
+        provider: 'xai',
+        label: 'Grok 4.6',
+        description: 'xAI frontier — coding, agentic tasks & knowledge work',
+        contextWindow: 500_000,
+        // "No text output limit" per docs.x.ai — conservative cap.
+        maxOutputTokens: 128_000,
+        supportsThinking: true,
+        thinkingBudgetTokens: 16_000,
+        thinkingConfigurable: true,
+        // reasoning_effort low|medium|high|xhigh, default high (xAI's model list
+        // `reasoningEffortOptions` for grok-4.6). Same surface as grok-4.7.
+        supportedEffortLevels: ['low', 'medium', 'high', 'xhigh'],
+        defaultEffortLevel: 'high',
+        // Multimodal input (text + images), text out.
+        supportsVision: true,
+        supportsPromptCaching: true,
+        supportsTools: true,
+        // <200K-token prompts; ≥200K bills 2× ($4/$1.00/$12 — tiering not modeled).
+        inputPricePerMTok: 2,
+        outputPricePerMTok: 6,
+        // xAI cached input billed at a flat $0.50/M, no write premium.
+        cacheReadPricePerMTok: 0.5,
+        cacheWritePricePerMTok: 2,
+        // Official (docs.x.ai/developers/grok-4-6, as of 2026-09-10): February 1, 2026.
+        knowledgeCutoff: '2026-02-01',
+        // Superseded by grok-4.7 on arrival: same general-purpose Grok line, same
+        // list price, window and effort levels, newer weights. Cataloged so a
+        // request naming grok-4.6 is priced; never offered in the picker.
+        deprecatedAt: '2026-09-30',
+        supersededBy: 'grok-4.7',
+    },
     {
         id: 'grok-4.5',
         provider: 'xai',
@@ -1470,6 +1544,10 @@ export const MODELS = [
         cacheWritePricePerMTok: 2,
         // Official (docs.x.ai): 2026-02-01.
         knowledgeCutoff: '2026-02-01',
+        // Superseded by grok-4.7: same general-purpose Grok line, same list price
+        // and window, newer weights. Stays priceable for saved selections.
+        deprecatedAt: '2026-09-30',
+        supersededBy: 'grok-4.7',
     },
     {
         id: 'grok-4.3',
@@ -1495,13 +1573,13 @@ export const MODELS = [
         cacheReadPricePerMTok: 0.2,
         cacheWritePricePerMTok: 1.25,
         knowledgeCutoff: '2025-12-01',
-        // Superseded by grok-4.5: the previous version of the same general-purpose
-        // Grok line, not a separately-named tier. It keeps a bigger window (1M vs
-        // 500K) and a lower price, which is why it was previously left selectable —
-        // but offering two generations of one family is exactly what the picker no
-        // longer does, and grok-4.5 is xAI's own recommendation. Stays priceable.
+        // Superseded by grok-4.7 (was grok-4.5 until 2026-09-30): the previous
+        // version of the same general-purpose Grok line, not a separately-named
+        // tier. It keeps a bigger window (1M vs 500K) and a lower price, which is
+        // why it was previously left selectable — but offering two generations of
+        // one family is exactly what the picker no longer does. Stays priceable.
         deprecatedAt: '2026-07-28',
-        supersededBy: 'grok-4.5',
+        supersededBy: 'grok-4.7',
     },
     {
         id: 'grok-build-0.1',

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@molecule/api-resource-ai-models",
-  "version": "1.11.0",
+  "version": "1.12.0",
   "description": "AI model catalog — server-side source of truth plus an authentication-gated discovery endpoint",
   "type": "module",
   "main": "dist/index.js",