slash-tokens 1.6.4 → 1.6.5

package/CHANGELOG.md

@@ -1,10 +1,20 @@
 # Changelog
 
+## [1.6.5] — The Fixed Deal Edition
+
+*2026-08-25*
+
+Solo $20 mailbox, 10% waived. Team $39 for the data.
+
+Rebuilt tarball. 1.6.4 packed a stale gitignored `dist/` (`bun test` runs `src/`). This tarball contains the live ladder: Grok **4.6** / **4.3**, GPT-5.6 **Sol / Terra / Luna**, Claude **Opus 5 / Sonnet 5 / Haiku 4.5**, Gemini **3.5 Flash-Lite**.
+
+No Team/Solo price change.
+
 ## [1.6.4] — The Fixed Deal Edition
 
 *2026-08-25*
 
-Live model ladder. Grok **4.6** / **4.3**, GPT-5.6 **Sol / Terra / Luna**, Claude **Opus 5 / Sonnet 5 / Haiku 4.5**, Gemini **3.5 Flash-Lite**. Old keys stay as aliases. Calibration factors carried from the 2026-08-23 corpus — not re-measured on the new IDs.
+Live model ladder in source. **Tarball packed stale `dist/`** — `bun test` never rebuilds it. Use **1.6.5**.
 
 No Team/Solo price change.
 

package/README.md

@@ -15,7 +15,7 @@ Know the cost before the call leaves your machine.
 Models change. Windows grow. Slash adapts — you keep building.
 Cheaper tokens haven't shrunk the bill — usage has.
 
-## v1.6.4 — The Fixed Deal Edition
+## v1.6.5 — The Fixed Deal Edition
 
 Solo $20 mailbox, 10% waived. Team $39 for the data.
 

package/dist/cli.js

@@ -52,14 +52,23 @@ function writeToMemory(content) {
 // src/slash.ts
 var WASM_INPUT_OFFSET2 = 4096;
 var CALIBRATION = {
+  "claude-opus-5": 2.05,
   "claude-opus": 2.05,
   "claude-opus-4.7": 2.05,
+  "claude-sonnet-5": 2.05,
   "claude-sonnet": 2.05,
+  "claude-haiku-4.5": 1.45,
   "claude-haiku": 1.45,
   "gemini-3.1-pro": 1.45,
+  "gemini-3.5-flash-lite": 1.45,
   "gemini-2.5-flash": 1.45,
+  "grok-4.6": 1.15,
+  "grok-4.3": 1.15,
   "grok-4.20": 1.15,
   "grok-4-1-fast": 1.15,
+  "gpt-5.6-sol": 1.15,
+  "gpt-5.6-terra": 1.15,
+  "gpt-5.6-luna": 1.15,
   "gpt-5.4": 1.15,
   "gpt-5.4-mini": 1.15,
   "gpt-5.4-nano": 1.15
@@ -91,12 +100,12 @@ var AI_PATTERNS = [
   { name: "Mistral", regex: /from\s+['"]@mistralai|MistralClient/g }
 ];
 var SDK_REPRESENTATIVE_MODEL = {
-  Anthropic: "claude-sonnet",
-  OpenAI: "gpt-5.4",
+  Anthropic: "claude-sonnet-5",
+  OpenAI: "gpt-5.6-sol",
   Gemini: "gemini-3.1-pro",
-  Grok: "grok-4.20"
+  Grok: "grok-4.6"
 };
-var UNKNOWN_SDK_REPRESENTATIVE_MODEL = "claude-sonnet";
+var UNKNOWN_SDK_REPRESENTATIVE_MODEL = "claude-sonnet-5";
 var SKIP_DIRS = new Set([
   "node_modules",
   ".git",
@@ -203,18 +212,61 @@ function scan(dir) {
 }
 
 // src/models.ts
+var OPUS = { input: 5, output: 25, context: 1e6 };
+var SONNET = { input: 2, output: 10, context: 1e6 };
+var HAIKU = { input: 1, output: 5, context: 200000 };
+var GROK_46 = {
+  input: 2,
+  output: 6,
+  context: 500000,
+  longContextThreshold: 200000,
+  longContextInput: 4,
+  longContextOutput: 12
+};
+var GROK_43 = {
+  input: 1.25,
+  output: 2.5,
+  context: 1e6,
+  longContextThreshold: 200000,
+  longContextInput: 2.5,
+  longContextOutput: 5
+};
+var GEMINI_PRO = {
+  input: 2,
+  output: 12,
+  context: 1e6,
+  longContextThreshold: 200000,
+  longContextInput: 4,
+  longContextOutput: 18
+};
+var GEMINI_FLASH = { input: 0.3, output: 2.5, context: 1e6 };
+var GPT_SOL = { input: 4, output: 20, context: 1050000 };
+var GPT_TERRA = { input: 2, output: 12, context: 1050000 };
+var GPT_LUNA = { input: 0.2, output: 1.2, context: 1050000 };
+var GPT_54 = { input: 2.5, output: 15, context: 1e6 };
+var GPT_54_MINI = { input: 0.75, output: 4.5, context: 128000 };
+var GPT_54_NANO = { input: 0.2, output: 1.25, context: 128000 };
 var MODELS = {
-  "claude-opus": { input: 5, output: 25, context: 1e6 },
-  "claude-opus-4.7": { input: 5, output: 25, context: 1e6 },
-  "claude-sonnet": { input: 2, output: 10, context: 1e6 },
-  "claude-haiku": { input: 1, output: 5, context: 200000 },
-  "grok-4.20": { input: 1.25, output: 2.5, context: 1e6, longContextThreshold: 200000, longContextInput: 2.5, longContextOutput: 5 },
-  "grok-4-1-fast": { input: 1.25, output: 2.5, context: 1e6, longContextThreshold: 200000, longContextInput: 2.5, longContextOutput: 5 },
-  "gemini-3.1-pro": { input: 2, output: 12, context: 1e6 },
-  "gemini-2.5-flash": { input: 0.3, output: 2.5, context: 1e6 },
-  "gpt-5.4": { input: 2.5, output: 15, context: 1e6 },
-  "gpt-5.4-mini": { input: 0.75, output: 4.5, context: 128000 },
-  "gpt-5.4-nano": { input: 0.2, output: 1.25, context: 128000 }
+  "claude-opus-5": { ...OPUS },
+  "claude-opus": { ...OPUS },
+  "claude-opus-4.7": { ...OPUS },
+  "claude-sonnet-5": { ...SONNET },
+  "claude-sonnet": { ...SONNET },
+  "claude-haiku-4.5": { ...HAIKU },
+  "claude-haiku": { ...HAIKU },
+  "grok-4.6": { ...GROK_46 },
+  "grok-4.3": { ...GROK_43 },
+  "grok-4.20": { ...GROK_43 },
+  "grok-4-1-fast": { ...GROK_43 },
+  "gemini-3.1-pro": { ...GEMINI_PRO },
+  "gemini-3.5-flash-lite": { ...GEMINI_FLASH },
+  "gemini-2.5-flash": { ...GEMINI_FLASH },
+  "gpt-5.6-sol": { ...GPT_SOL },
+  "gpt-5.6-terra": { ...GPT_TERRA },
+  "gpt-5.6-luna": { ...GPT_LUNA },
+  "gpt-5.4": { ...GPT_54 },
+  "gpt-5.4-mini": { ...GPT_54_MINI },
+  "gpt-5.4-nano": { ...GPT_54_NANO }
 };
 function getModel(name) {
   return MODELS[name] || MODELS[name.toLowerCase()];

package/dist/intercept.js

@@ -5,16 +5,25 @@ import { PROVIDER_MODELS } from './providers.js';
 // Reverse lookup: model name → provider model names in the API
 // (what to put back in the request body)
 const MODEL_API_NAMES = {
+    'claude-opus-5': 'claude-opus-5',
     'claude-opus': 'claude-opus-5',
     'claude-opus-4.7': 'claude-opus-4-7',
+    'claude-sonnet-5': 'claude-sonnet-5',
     'claude-sonnet': 'claude-sonnet-5',
+    'claude-haiku-4.5': 'claude-haiku-4-5-20251001',
     'claude-haiku': 'claude-haiku-4-5-20251001',
+    'gpt-5.6-sol': 'gpt-5.6-sol',
+    'gpt-5.6-terra': 'gpt-5.6-terra',
+    'gpt-5.6-luna': 'gpt-5.6-luna',
     'gpt-5.4': 'gpt-5.4',
     'gpt-5.4-mini': 'gpt-5.4-mini',
     'gpt-5.4-nano': 'gpt-5.4-nano',
+    'grok-4.6': 'grok-4.6',
+    'grok-4.3': 'grok-4.3',
     'grok-4.20': 'grok-4.20-0309-non-reasoning',
     'grok-4-1-fast': 'grok-4.3',
     'gemini-3.1-pro': 'gemini-pro-latest',
+    'gemini-3.5-flash-lite': 'gemini-3.5-flash-lite',
     'gemini-2.5-flash': 'gemini-flash-latest',
 };
 // AI API endpoint detection
@@ -27,7 +36,7 @@ const AI_ENDPOINTS = [
     {
         pattern: /api\.openai\.com/,
         provider: 'OpenAI',
-        modelExtractor: (body) => body?.model || 'gpt-5.4',
+        modelExtractor: (body) => body?.model || 'gpt-5.6-sol',
     },
     {
         pattern: /generativelanguage\.googleapis\.com/,
@@ -35,13 +44,13 @@ const AI_ENDPOINTS = [
         modelExtractor: (_body, url) => {
             // Model is in the URL path: /v1beta/models/gemini-2.0-flash:generateContent
             const match = url?.match(/\/models\/([^/:]+)/);
-            return match ? match[1] : 'gemini-2.5-flash';
+            return match ? match[1] : 'gemini-3.5-flash-lite';
         },
     },
     {
         pattern: /api\.x\.ai/,
         provider: 'xAI',
-        modelExtractor: (body) => body?.model || 'grok-4.20',
+        modelExtractor: (body) => body?.model || 'grok-4.6',
     },
 ];
 // Normalize model names to our pricing table keys.
@@ -55,33 +64,54 @@ const AI_ENDPOINTS = [
 // to a falsely-safe-looking value is the dangerous case).
 export function normalizeModel(raw) {
     const lower = raw.toLowerCase();
-    // Anthropic — 4.7 before generic opus check, same order as slash-models.ts
+    // Anthropic — specific versions before generic family
     if (lower.includes('opus') && (lower.includes('4-7') || lower.includes('4.7')))
         return 'claude-opus-4.7';
+    if (lower.includes('opus-5'))
+        return 'claude-opus-5';
     if (lower.includes('opus'))
         return 'claude-opus';
+    if (lower.includes('sonnet-5'))
+        return 'claude-sonnet-5';
     if (lower.includes('sonnet'))
         return 'claude-sonnet';
+    if (lower.includes('haiku') && (lower.includes('4.5') || lower.includes('4-5')))
+        return 'claude-haiku-4.5';
     if (lower.includes('haiku'))
         return 'claude-haiku';
-    // xAI
+    // xAI — 4.6 / 4.3 / 4.20 before generic grok
+    if (lower.includes('grok') && (lower.includes('4.6') || lower.includes('4-6')))
+        return 'grok-4.6';
+    if (lower.includes('grok') && (lower.includes('4.3') || lower.includes('4-3')))
+        return 'grok-4.3';
+    if (lower.includes('grok') && (lower.includes('4.20') || lower.includes('4-20')))
+        return 'grok-4.20';
     if (lower.includes('grok') && lower.includes('fast'))
         return 'grok-4-1-fast';
     if (lower.includes('grok'))
-        return 'grok-4.20';
+        return 'grok-4.6';
     // Google
     if (lower.includes('gemini') && lower.includes('pro'))
         return 'gemini-3.1-pro';
-    if (lower.includes('gemini'))
+    if (lower.includes('gemini') && lower.includes('3.5') && lower.includes('lite'))
+        return 'gemini-3.5-flash-lite';
+    if (lower.includes('gemini') && lower.includes('2.5'))
         return 'gemini-2.5-flash';
-    // OpenAI — current (5.4 family)
+    if (lower.includes('gemini'))
+        return 'gemini-3.5-flash-lite';
+    // OpenAI — 5.6 then 5.4 then legacy
+    if (lower.includes('5.6') && lower.includes('luna'))
+        return 'gpt-5.6-luna';
+    if (lower.includes('5.6') && lower.includes('terra'))
+        return 'gpt-5.6-terra';
+    if (lower.includes('5.6'))
+        return 'gpt-5.6-sol';
     if (lower.includes('5.4') && lower.includes('nano'))
         return 'gpt-5.4-nano';
     if (lower.includes('5.4') && lower.includes('mini'))
         return 'gpt-5.4-mini';
     if (lower.includes('5.4'))
         return 'gpt-5.4';
-    // OpenAI — legacy model names → map to closest current equivalent
     if (lower.includes('o1-mini') || lower.includes('o1_mini'))
         return 'gpt-5.4-mini';
     if (lower.includes('o1'))

package/dist/models.js

@@ -1,44 +1,60 @@
-// Pricing as of April 2026 — USD per million tokens
-// Claude Sonnet 5 pricing corrected 2026-08-23: was hardcoded at $3.00/
-// $15.00 (the previously-scheduled Sept 1, 2026 increase), but Anthropic's
-// own pricing page confirms that increase will NOT occur — $2.00/$10.00
-// (the introductory rate) is now the permanent standard price. Found via
-// a full-codebase review that independently re-verified every provider's
-// live pricing, not just the model IDs already fixed that day.
-// xAI pricing re-derived 2026-08-23: grok-4.20 and grok-4-1-fast (the literal
-// API IDs, not just the generic keys below) were both fully retired — not
-// just old snapshots, they 404 on the live API. Current lineup has no cheap
-// tier at all; grok-4.20 (generic) now targets grok-4.20-0309-non-reasoning
-// and grok-4-1-fast (generic) targets grok-4.3, both $1.25/$2.50/1M — see
-// intercept.ts MODEL_API_NAMES. There is currently no xAI model cheaper than
-// $1.25/M input, so routing between these two generic keys yields zero
-// savings (findCheapestRoute requires strictly cheaper — this is honest,
-// not a bug: the old $0.20/M "fast" tier no longer exists).
+const OPUS = { input: 5.00, output: 25.00, context: 1000000 };
+const SONNET = { input: 2.00, output: 10.00, context: 1000000 };
+const HAIKU = { input: 1.00, output: 5.00, context: 200000 };
+const GROK_46 = {
+    input: 2.00, output: 6.00, context: 500000,
+    longContextThreshold: 200000, longContextInput: 4.00, longContextOutput: 12.00,
+};
+const GROK_43 = {
+    input: 1.25, output: 2.50, context: 1000000,
+    longContextThreshold: 200000, longContextInput: 2.50, longContextOutput: 5.00,
+};
+const GEMINI_PRO = {
+    input: 2.00, output: 12.00, context: 1000000,
+    longContextThreshold: 200000, longContextInput: 4.00, longContextOutput: 18.00,
+};
+const GEMINI_FLASH = { input: 0.30, output: 2.50, context: 1000000 };
+const GPT_SOL = { input: 4.00, output: 20.00, context: 1050000 };
+const GPT_TERRA = { input: 2.00, output: 12.00, context: 1050000 };
+const GPT_LUNA = { input: 0.20, output: 1.20, context: 1050000 };
+const GPT_54 = { input: 2.50, output: 15.00, context: 1000000 };
+const GPT_54_MINI = { input: 0.75, output: 4.50, context: 128000 };
+const GPT_54_NANO = { input: 0.20, output: 1.25, context: 128000 };
+// Pricing as of 2026-08-25 — USD per million tokens.
+// First-party: platform.claude.com/docs/en/about-claude/pricing
+//              developers.openai.com/api/docs/models
+//              docs.x.ai/developers/models
+//              ai.google.dev/gemini-api/docs/pricing
+// Old keys stay as aliases so existing call sites don't throw.
 export const MODELS = {
-    // Anthropic
-    'claude-opus': { input: 5.00, output: 25.00, context: 1000000 },
-    'claude-opus-4.7': { input: 5.00, output: 25.00, context: 1000000 },
-    'claude-sonnet': { input: 2.00, output: 10.00, context: 1000000 },
-    'claude-haiku': { input: 1.00, output: 5.00, context: 200000 },
-    // xAI
-    'grok-4.20': { input: 1.25, output: 2.50, context: 1000000, longContextThreshold: 200000, longContextInput: 2.50, longContextOutput: 5.00 },
-    'grok-4-1-fast': { input: 1.25, output: 2.50, context: 1000000, longContextThreshold: 200000, longContextInput: 2.50, longContextOutput: 5.00 },
+    // Anthropic — live names + generic aliases (same rates)
+    'claude-opus-5': { ...OPUS },
+    'claude-opus': { ...OPUS },
+    'claude-opus-4.7': { ...OPUS },
+    'claude-sonnet-5': { ...SONNET },
+    'claude-sonnet': { ...SONNET },
+    'claude-haiku-4.5': { ...HAIKU },
+    'claude-haiku': { ...HAIKU },
+    // xAI — flagship 4.6, cheap same-provider 4.3. 4.20 / fast are aliases.
+    'grok-4.6': { ...GROK_46 },
+    'grok-4.3': { ...GROK_43 },
+    'grok-4.20': { ...GROK_43 },
+    'grok-4-1-fast': { ...GROK_43 },
     // Google
-    'gemini-3.1-pro': { input: 2.00, output: 12.00, context: 1000000 },
-    'gemini-2.5-flash': { input: 0.30, output: 2.50, context: 1000000 },
-    // OpenAI
-    'gpt-5.4': { input: 2.50, output: 15.00, context: 1000000 },
-    'gpt-5.4-mini': { input: 0.75, output: 4.50, context: 128000 },
-    'gpt-5.4-nano': { input: 0.20, output: 1.25, context: 128000 },
+    'gemini-3.1-pro': { ...GEMINI_PRO },
+    'gemini-3.5-flash-lite': { ...GEMINI_FLASH },
+    'gemini-2.5-flash': { ...GEMINI_FLASH },
+    // OpenAI — live 5.6 ladder. 5.4 family kept as aliases (old prices).
+    'gpt-5.6-sol': { ...GPT_SOL },
+    'gpt-5.6-terra': { ...GPT_TERRA },
+    'gpt-5.6-luna': { ...GPT_LUNA },
+    'gpt-5.4': { ...GPT_54 },
+    'gpt-5.4-mini': { ...GPT_54_MINI },
+    'gpt-5.4-nano': { ...GPT_54_NANO },
 };
 export function getModel(name) {
     return MODELS[name] || MODELS[name.toLowerCase()];
 }
-// Resolve the actual billable rate for a given token count — applies the
-// long-context tier above if the model has one and tokens cross it.
-// preflight()/preflightRoute() should go through this, not read
-// .input/.output directly, or a Grok call over 200K tokens gets silently
-// under-costed at the base rate.
 export function effectiveRate(tokens, info) {
     if (info.longContextThreshold !== undefined && tokens > info.longContextThreshold) {
         return {

package/dist/patterns.d.ts

@@ -4,6 +4,6 @@ export interface Pattern {
 }
 export declare const AI_PATTERNS: Pattern[];
 export declare const SDK_REPRESENTATIVE_MODEL: Record<string, string>;
-export declare const UNKNOWN_SDK_REPRESENTATIVE_MODEL = "claude-sonnet";
+export declare const UNKNOWN_SDK_REPRESENTATIVE_MODEL = "claude-sonnet-5";
 export declare const SKIP_DIRS: Set<string>;
 export declare const SCAN_EXTENSIONS: Set<string>;

package/dist/patterns.js

@@ -34,17 +34,17 @@ export const AI_PATTERNS = [
 // know the exact model), but a real per-provider one instead of a single
 // guess applied to everyone.
 export const SDK_REPRESENTATIVE_MODEL = {
-    'Anthropic': 'claude-sonnet',
-    'OpenAI': 'gpt-5.4',
+    'Anthropic': 'claude-sonnet-5',
+    'OpenAI': 'gpt-5.6-sol',
     'Gemini': 'gemini-3.1-pro',
-    'Grok': 'grok-4.20',
+    'Grok': 'grok-4.6',
 };
 // Fallback for SDKs that don't map to one specific provider (Vercel AI,
 // LangChain, and Bedrock can all wrap any underlying provider; raw
 // fetch-to-AI-endpoint and Cohere/Mistral have no pricing data in MODELS
 // at all). claude-sonnet is used as a documented, honest middle-of-the-
 // road placeholder — not a claim about which model is actually running.
-export const UNKNOWN_SDK_REPRESENTATIVE_MODEL = 'claude-sonnet';
+export const UNKNOWN_SDK_REPRESENTATIVE_MODEL = 'claude-sonnet-5';
 export const SKIP_DIRS = new Set([
     'node_modules', '.git', 'dist', 'build', '.next', '.nuxt', '.svelte-kit',
     'coverage', '.turbo', '.cache', '__pycache__', '.venv', 'venv',

package/dist/providers.d.ts

@@ -1,26 +1,11 @@
 /**
  * Provider groups — single source of truth.
  *
- * Slash routing is always SAME-PROVIDER. Opus → Haiku, GPT-5.4 → Nano,
- * Grok-4.20 → Fast, Gemini 3.1 Pro → 2.5 Flash. Never cross-provider.
+ * Slash routing is always SAME-PROVIDER. Opus 5 → Haiku 4.5, Sol → Luna,
+ * Grok 4.6 → 4.3, Gemini 3.1 Pro → 3.5 Flash-Lite. Never cross-provider.
  *
- * This file is the canonical provider mapping. Both `intercept.ts`
- * (runtime fetch patching) and `preflight.ts` (analysis + routing
- * prediction) consume from here so they can never drift.
- *
- * TEST-NOTE: Whenever a new model is added, it MUST appear in exactly one
- * provider group below. A model missing from here will:
- *   - Never be a routing target from `findCheapestRoute` / `preflightRoute`
- *   - Still appear in `preflight().options` (which is cross-provider analysis)
- *   That mismatch is by design — see preflight.ts semantics.
+ * Order inside a group matters when two models share a price: the first
+ * strictly-cheaper hit wins (findCheapestRoute / preflightRoute).
  */
 export declare const PROVIDER_MODELS: Record<string, string[]>;
-/**
- * Identify a model's provider from its canonical key.
- *
- * TEST-NOTE: This function MUST return a non-null provider for every model
- * present in `MODELS` (from models.ts). If `MODELS` adds a model without
- * adding it to `PROVIDER_MODELS`, this returns null and routing is disabled
- * for that model. Silent skip.
- */
 export declare function providerOf(model: string): string | null;

package/dist/providers.js

@@ -1,33 +1,26 @@
 /**
  * Provider groups — single source of truth.
  *
- * Slash routing is always SAME-PROVIDER. Opus → Haiku, GPT-5.4 → Nano,
- * Grok-4.20 → Fast, Gemini 3.1 Pro → 2.5 Flash. Never cross-provider.
+ * Slash routing is always SAME-PROVIDER. Opus 5 → Haiku 4.5, Sol → Luna,
+ * Grok 4.6 → 4.3, Gemini 3.1 Pro → 3.5 Flash-Lite. Never cross-provider.
  *
- * This file is the canonical provider mapping. Both `intercept.ts`
- * (runtime fetch patching) and `preflight.ts` (analysis + routing
- * prediction) consume from here so they can never drift.
- *
- * TEST-NOTE: Whenever a new model is added, it MUST appear in exactly one
- * provider group below. A model missing from here will:
- *   - Never be a routing target from `findCheapestRoute` / `preflightRoute`
- *   - Still appear in `preflight().options` (which is cross-provider analysis)
- *   That mismatch is by design — see preflight.ts semantics.
+ * Order inside a group matters when two models share a price: the first
+ * strictly-cheaper hit wins (findCheapestRoute / preflightRoute).
  */
 export const PROVIDER_MODELS = {
-    Anthropic: ['claude-opus', 'claude-opus-4.7', 'claude-sonnet', 'claude-haiku'],
-    OpenAI: ['gpt-5.4', 'gpt-5.4-mini', 'gpt-5.4-nano'],
-    xAI: ['grok-4.20', 'grok-4-1-fast'],
-    Google: ['gemini-3.1-pro', 'gemini-2.5-flash'],
+    Anthropic: [
+        'claude-opus-5', 'claude-opus', 'claude-opus-4.7',
+        'claude-sonnet-5', 'claude-sonnet',
+        'claude-haiku', 'claude-haiku-4.5',
+    ],
+    OpenAI: [
+        'gpt-5.6-sol', 'gpt-5.6-terra',
+        'gpt-5.4', 'gpt-5.4-mini', 'gpt-5.4-nano',
+        'gpt-5.6-luna',
+    ],
+    xAI: ['grok-4.6', 'grok-4.3', 'grok-4.20', 'grok-4-1-fast'],
+    Google: ['gemini-3.1-pro', 'gemini-3.5-flash-lite', 'gemini-2.5-flash'],
 };
-/**
- * Identify a model's provider from its canonical key.
- *
- * TEST-NOTE: This function MUST return a non-null provider for every model
- * present in `MODELS` (from models.ts). If `MODELS` adds a model without
- * adding it to `PROVIDER_MODELS`, this returns null and routing is disabled
- * for that model. Silent skip.
- */
 export function providerOf(model) {
     for (const [provider, models] of Object.entries(PROVIDER_MODELS)) {
         if (models.includes(model))

package/dist/slash.js

@@ -65,7 +65,10 @@ const WASM_INPUT_OFFSET = 4096;
  *             DEFAULT_UNKNOWN_MODEL_FACTOR (1.85), a ~60% larger correction
  *             than it needed. Still safe either way (1.85 > required
  *             minimum), just needlessly inflated for real GPT users.
- * grok-4.20 / grok-4-1-fast: 1.15 — re-verified 2026-08-23 against the
+ * grok-4.6 / grok-4.3 (and aliases grok-4.20 / grok-4-1-fast): 1.15 —
+ *             4.6/4.3 IDs added 2026-08-25; factor CARRIED from the
+ *             2026-08-23 corpus, not re-measured on the new wire IDs.
+ * grok-4.20 / grok-4-1-fast (original): 1.15 — re-verified 2026-08-23 against the
  *             29-sample corpus (20 new samples: more languages, Spanish/
  *             Japanese prose, more JSON shapes) and UNCHANGED — same
  *             worst case (technical-docs prose, ratio 0.928) as the
@@ -91,14 +94,23 @@ const WASM_INPUT_OFFSET = 4096;
  * Slash must NEVER under-report. Over-reporting is safe (go/no-go only).
  */
 const CALIBRATION = {
+    'claude-opus-5': 2.05,
     'claude-opus': 2.05,
     'claude-opus-4.7': 2.05,
+    'claude-sonnet-5': 2.05,
     'claude-sonnet': 2.05,
+    'claude-haiku-4.5': 1.45,
     'claude-haiku': 1.45,
     'gemini-3.1-pro': 1.45,
+    'gemini-3.5-flash-lite': 1.45,
     'gemini-2.5-flash': 1.45,
+    'grok-4.6': 1.15,
+    'grok-4.3': 1.15,
     'grok-4.20': 1.15,
     'grok-4-1-fast': 1.15,
+    'gpt-5.6-sol': 1.15,
+    'gpt-5.6-terra': 1.15,
+    'gpt-5.6-luna': 1.15,
     'gpt-5.4': 1.15,
     'gpt-5.4-mini': 1.15,
     'gpt-5.4-nano': 1.15,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "slash-tokens",
-  "version": "1.6.4",
+  "version": "1.6.5",
   "description": "Token Optimization for Context Engineers. 4.8 KB WASM. Sub-millisecond. Zero dependencies.",
   "main": "dist/index.js",
   "module": "dist/index.js",