@aliou/pi-neuralwatt 0.7.3 → 0.7.6

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@aliou/pi-neuralwatt",
-  "version": "0.7.3",
+  "version": "0.7.6",
   "license": "MIT",
   "type": "module",
   "private": false,

package/src/extensions/provider/index.ts

@@ -46,6 +46,11 @@ function registerNeuralwattProvider(
   const { includeLegacyModelIds, includeHiddenModels } =
     configLoader.getConfig();
 
+  const publicModels = getNeuralwattModels({ includeLegacyModelIds });
+  const resolvedHiddenModels = includeHiddenModels
+    ? dedupeHiddenModels(hiddenModels, publicModels)
+    : [];
+
   const config: Parameters<ExtensionAPI["registerProvider"]>[1] = {
     baseUrl: "https://api.neuralwatt.com/v1",
     apiKey: "$NEURALWATT_API_KEY",
@@ -55,10 +60,7 @@ function registerNeuralwattProvider(
       Referer: "https://pi.dev",
       "X-Title": "npm:@aliou/pi-neuralwatt",
     },
-    models: [
-      ...getNeuralwattModels({ includeLegacyModelIds }),
-      ...(includeHiddenModels ? hiddenModels : []),
-    ],
+    models: [...publicModels, ...resolvedHiddenModels],
   };
 
   const provider = getApiProvider("openai-completions");
@@ -73,6 +75,23 @@ function registerNeuralwattProvider(
   pi.registerProvider("neuralwatt", config);
 }
 
+/**
+ * Drop any hidden model whose ID collides with a public or legacy model.
+ *
+ * Models can graduate from hidden (authenticated /v1/models only) to public
+ * (unauthenticated list). When that happens, a stale on-disk cache may still
+ * list the now-public ID, which would register it twice and make Pi treat the
+ * scoped model as ambiguous ("No models match pattern"). Dedupe against the
+ * public list so a stale cache can never shadow a public model.
+ */
+function dedupeHiddenModels(
+  hiddenModels: ProviderModelConfig[],
+  publicModels: ProviderModelConfig[],
+): ProviderModelConfig[] {
+  const publicIds = new Set(publicModels.map((m) => m.id));
+  return hiddenModels.filter((m) => !publicIds.has(m.id));
+}
+
 export default async function (pi: ExtensionAPI) {
   await configLoader.load();
 
@@ -254,14 +273,14 @@ export default async function (pi: ExtensionAPI) {
         ctx.modelRegistry.authStorage,
         hiddenModelsAbort.signal,
       );
-      // Persist for the next startup so scoped models resolve without warnings
-      // on Pi's subsequent launches.
-      if (fetched.length > 0) {
-        hiddenModels = fetched;
-        await writeHiddenModelsCache(hiddenModels);
-        if (!hiddenModelsAbort.signal.aborted) {
-          registerNeuralwattProvider(pi, handleSseQuota, hiddenModels);
-        }
+      // Persist for the next startup so scoped models resolve without
+      // warnings on Pi's subsequent launches. Always write the cache (even
+      // when empty) and re-register, so graduated or removed hidden models
+      // are purged from both the cache and the provider's model list.
+      hiddenModels = fetched;
+      await writeHiddenModelsCache(hiddenModels);
+      if (!hiddenModelsAbort.signal.aborted) {
+        registerNeuralwattProvider(pi, handleSseQuota, hiddenModels);
       }
     }
 

package/src/extensions/provider/models/legacy.ts

@@ -7,6 +7,8 @@ export const LEGACY_MODEL_ALIAS_MAP = {
   "glm-5.1": "glm-5.2",
   "glm-5.1-fast": "glm-5.2-fast",
   "zai-org/GLM-5.1-FP8": "glm-5.2",
+  "moonshotai/Kimi-K2.5": "kimi-k2.6",
+  "kimi-k2.5-fast": "kimi-k2.6-fast",
   "moonshotai/Kimi-K2.6": "kimi-k2.6",
   "Qwen/Qwen3.5-397B-A17B-FP8": "qwen3.5-397b",
   "Qwen/Qwen3.6-35B-A3B": "qwen3.6-35b",

package/src/extensions/provider/models/public-models.ts

@@ -18,6 +18,7 @@ export const NEURALWATT_MODELS: ProviderModelConfig[] = [
     contextWindow: 1048560,
     maxTokens: 65536,
     thinkingLevelMap: {
+      off: "none",
       minimal: null,
       low: null,
       medium: null,
@@ -64,6 +65,7 @@ export const NEURALWATT_MODELS: ProviderModelConfig[] = [
     contextWindow: 199984,
     maxTokens: 65536,
     thinkingLevelMap: {
+      off: "none",
       minimal: null,
       low: null,
       medium: null,
@@ -95,52 +97,6 @@ export const NEURALWATT_MODELS: ProviderModelConfig[] = [
       maxTokensField: "max_tokens",
     },
   },
-  // Kimi K2.5 - MoonshotAI
-  {
-    id: "moonshotai/Kimi-K2.5",
-    name: "Kimi K2.5",
-    reasoning: true,
-    input: ["text", "image"],
-    cost: {
-      input: 0.52,
-      output: 2.59,
-      cacheRead: 0.13,
-      cacheWrite: 0,
-    },
-    contextWindow: 262128,
-    maxTokens: 65536,
-    thinkingLevelMap: {
-      minimal: null,
-      low: null,
-      medium: "medium",
-      high: null,
-      xhigh: null,
-    },
-    compat: {
-      supportsDeveloperRole: false,
-      maxTokensField: "max_tokens",
-      requiresReasoningContentOnAssistantMessages: true,
-    },
-  },
-  // Kimi K2.5 Fast - MoonshotAI
-  {
-    id: "kimi-k2.5-fast",
-    name: "Kimi K2.5 Fast",
-    reasoning: false,
-    input: ["text", "image"],
-    cost: {
-      input: 0.52,
-      output: 2.59,
-      cacheRead: 0.13,
-      cacheWrite: 0,
-    },
-    contextWindow: 262128,
-    maxTokens: 65536,
-    compat: {
-      supportsDeveloperRole: false,
-      maxTokensField: "max_tokens",
-    },
-  },
   // Kimi K2.6 - MoonshotAI
   {
     id: "kimi-k2.6",
@@ -288,6 +244,136 @@ export const NEURALWATT_MODELS: ProviderModelConfig[] = [
       requiresReasoningContentOnAssistantMessages: true,
     },
   },
+  // GLM-5.2 Short Fast Flex - ZhipuAI (flex variant, reasoning disabled)
+  {
+    id: "glm-5.2-short-fast-flex",
+    name: "GLM-5.2 (short, fast, flex)",
+    reasoning: false,
+    input: ["text"],
+    cost: {
+      input: 1.45,
+      output: 4.5,
+      cacheRead: 0.3625,
+      cacheWrite: 0,
+    },
+    contextWindow: 199984,
+    maxTokens: 65536,
+    compat: {
+      supportsDeveloperRole: false,
+      maxTokensField: "max_tokens",
+    },
+  },
+  // GLM-5.2 Short Flex - ZhipuAI (flex variant)
+  {
+    id: "glm-5.2-short-flex",
+    name: "GLM-5.2 (short, flex)",
+    reasoning: true,
+    input: ["text"],
+    cost: {
+      input: 1.45,
+      output: 4.5,
+      cacheRead: 0.3625,
+      cacheWrite: 0,
+    },
+    contextWindow: 199984,
+    maxTokens: 65536,
+    thinkingLevelMap: {
+      off: "none",
+      minimal: null,
+      low: null,
+      medium: null,
+      high: "high",
+      xhigh: "max",
+    },
+    compat: {
+      supportsDeveloperRole: false,
+      maxTokensField: "max_tokens",
+      requiresReasoningContentOnAssistantMessages: true,
+    },
+  },
+  // Kimi K2.6 Flex - MoonshotAI (flex variant)
+  {
+    id: "kimi-k2.6-flex",
+    name: "Kimi K2.6 (flex)",
+    reasoning: true,
+    input: ["text", "image"],
+    cost: {
+      input: 0.69,
+      output: 3.22,
+      cacheRead: 0.1725,
+      cacheWrite: 0,
+    },
+    contextWindow: 262128,
+    maxTokens: 65536,
+    thinkingLevelMap: {
+      minimal: null,
+      low: null,
+      medium: "medium",
+      high: null,
+      xhigh: null,
+    },
+    compat: {
+      supportsDeveloperRole: false,
+      maxTokensField: "max_tokens",
+      requiresReasoningContentOnAssistantMessages: true,
+    },
+  },
+  // Kimi K2.7 Code Flex - MoonshotAI (flex variant)
+  {
+    id: "kimi-k2.7-code-flex",
+    name: "Kimi K2.7 Code (flex)",
+    reasoning: true,
+    input: ["text", "image"],
+    cost: {
+      input: 0.95,
+      output: 4.0,
+      cacheRead: 0.2375,
+      cacheWrite: 0,
+    },
+    contextWindow: 262128,
+    maxTokens: 65536,
+    thinkingLevelMap: {
+      off: null,
+      minimal: null,
+      low: null,
+      medium: "medium",
+      high: null,
+      xhigh: null,
+    },
+    compat: {
+      supportsDeveloperRole: false,
+      maxTokensField: "max_tokens",
+      requiresReasoningContentOnAssistantMessages: true,
+    },
+  },
+  // GLM-5.2 Flex - ZhipuAI (flex variant)
+  {
+    id: "glm-5.2-flex",
+    name: "GLM-5.2 (flex)",
+    reasoning: true,
+    input: ["text"],
+    cost: {
+      input: 1.45,
+      output: 4.5,
+      cacheRead: 0.3625,
+      cacheWrite: 0,
+    },
+    contextWindow: 1048560,
+    maxTokens: 65536,
+    thinkingLevelMap: {
+      off: "none",
+      minimal: null,
+      low: null,
+      medium: null,
+      high: "high",
+      xhigh: "max",
+    },
+    compat: {
+      supportsDeveloperRole: false,
+      maxTokensField: "max_tokens",
+      requiresReasoningContentOnAssistantMessages: true,
+    },
+  },
   // Qwen3.6 35B Fast - Qwen
   {
     id: "qwen3.6-35b-fast",