Preparation to hide free MiniMax and deprecated auto models (#1166)

Author: chrarnoldus

src/app/api/openrouter/[...path]/route.ts

@@ -36,6 +36,7 @@ import {
   temporarilyUnavailableResponse,
   usageLimitExceededResponse,
   wrapInSafeNextResponse,
+  forbiddenFreeModelResponse,
 } from '@/lib/llm-proxy-helpers';
 import { getBalanceAndOrgSettings } from '@/lib/organizations/organization-usage';
 import { ENABLE_TOOL_REPAIR, repairTools } from '@/lib/tool-calling';
@@ -64,7 +65,7 @@ import {
 import { handleRequestLogging } from '@/lib/handleRequestLogging';
 import { customLlmRequest } from '@/lib/custom-llm/customLlmRequest';
 import { normalizeModelId } from '@/lib/model-utils';
-import { isRateLimitedToDeath } from '@/lib/rate-limited-models';
+import { isForbiddenFreeModel } from '@/lib/forbidden-free-models';
 import { isActiveReviewPromo } from '@/lib/code-reviews/core/constants';
 import { applyResolvedAutoModel, isKiloAutoModel } from '@/lib/kilo-auto-model';
 import { fixOpenCodeDuplicateReasoning } from '@/lib/providers/fixOpenCodeDuplicateReasoning';
@@ -318,13 +319,16 @@ export async function POST(request: NextRequest): Promise<NextResponseType<unkno
     return temporarilyUnavailableResponse();
   }
 
-  if (isDeadFreeModel(originalModelIdLowerCased)) {
-    console.warn(`User requested discontinued free model ${originalModelIdLowerCased}; rejecting.`);
-    return alphaPeriodEndedResponse();
-  }
-
-  if (isRateLimitedToDeath(originalModelIdLowerCased)) {
-    return modelDoesNotExistResponse();
+  if (
+    isDeadFreeModel(originalModelIdLowerCased) ||
+    (!autoModel && isForbiddenFreeModel(originalModelIdLowerCased))
+  ) {
+    console.warn(`User requested forbidden free model ${originalModelIdLowerCased}; rejecting.`);
+    if (isRooCodeBasedClient(fraudHeaders)) {
+      return alphaPeriodEndedResponse();
+    } else {
+      return forbiddenFreeModelResponse();
+    }
   }
 
   // Extract properties for usage context

src/lib/rate-limited-models.ts src/lib/forbidden-free-models.tssrc/lib/rate-limited-models.ts renamed to src/lib/forbidden-free-models.ts

@@ -5,7 +5,7 @@
  * 429s.
  */
 
-const rateLimitedToDeathModelIds: ReadonlySet<string> = new Set([
+const forbiddenFreeModelIds: ReadonlySet<string> = new Set([
   'arcee-ai/trinity-mini:free',
   'cognitivecomputations/dolphin-mistral-24b-venice-edition:free',
   'deepseek/deepseek-r1-0528:free',
@@ -32,6 +32,6 @@ const rateLimitedToDeathModelIds: ReadonlySet<string> = new Set([
   'z-ai/glm-4.5-air:free',
 ]);
 
-export function isRateLimitedToDeath(modelId: string): boolean {
-  return rateLimitedToDeathModelIds.has(modelId);
+export function isForbiddenFreeModel(modelId: string): boolean {
+  return forbiddenFreeModelIds.has(modelId);
 }

src/lib/llm-proxy-helpers.ts

@@ -31,6 +31,7 @@ import type {
   PromptInfo,
 } from '@/lib/processUsage.types';
 import { getMaxTokens } from '@/lib/providers/openrouter/request-helpers';
+import { KILO_AUTO_BALANCED_MODEL, KILO_AUTO_FREE_MODEL } from '@/lib/kilo-auto-model';
 
 // FIM suffix markers for tracking purposes - used to wrap suffix in a fake system prompt format
 // This allows FIM requests to be tracked consistently with chat requests
@@ -183,6 +184,11 @@ export function modelNotAllowedResponse() {
   );
 }
 
+export function forbiddenFreeModelResponse() {
+  const error = `This is not a free model. Please use ${KILO_AUTO_BALANCED_MODEL.id} for affordable inference or ${KILO_AUTO_FREE_MODEL.id} for limited free inference.`;
+  return NextResponse.json({ error, message: error }, { status: 404 });
+}
+
 export function modelDoesNotExistResponse() {
   return NextResponse.json(
     {

src/lib/providers/morph.ts

@@ -7,7 +7,7 @@ export const morph_warp_grep_free_model: KiloFreeModel = {
     'A code search subagent that finds relevant code in a separate context window — no embeddings, no indexing.',
   context_length: 256000,
   max_completion_tokens: 32000,
-  status: 'public',
+  status: 'hidden',
   flags: [],
   gateway: 'morph',
   internal_id: 'morph-warp-grep-v2',

src/lib/providers/openrouter/index.ts

@@ -8,7 +8,7 @@ import {
 import { errorExceptInTest } from '@/lib/utils.server';
 import { captureException, captureMessage } from '@sentry/nextjs';
 import { convertFromKiloModel } from '@/lib/providers/kilo-free-model';
-import { isRateLimitedToDeath } from '@/lib/rate-limited-models';
+import { isForbiddenFreeModel } from '@/lib/forbidden-free-models';
 import {
   getModelSettings,
   getOpenCodeSettings,
@@ -60,8 +60,7 @@ function enhancedModelList(models: OpenRouterModel[]) {
   const enhancedModels = models
     .filter(
       (model: OpenRouterModel) =>
-        !kiloFreeModels.some(m => m.public_id === model.id && m.status !== 'disabled') &&
-        !isRateLimitedToDeath(model.id)
+        !kiloFreeModels.some(m => m.public_id === model.id) && !isForbiddenFreeModel(model.id)
     )
     .concat(
       kiloFreeModels.filter(m => m.status === 'public').map(model => convertFromKiloModel(model))

src/tests/openrouter-models-sorting.approved.json

@@ -741,47 +741,6 @@
       "default_parameters": {},
       "isFree": true
     },
-    {
-      "id": "morph-warp-grep-v2",
-      "canonical_slug": "morph-warp-grep-v2",
-      "hugging_face_id": "",
-      "name": "Morph: WarpGrep V2",
-      "created": 1756238927,
-      "description": "A code search subagent that finds relevant code in a separate context window — no embeddings, no indexing.",
-      "context_length": 256000,
-      "architecture": {
-        "modality": "text->text",
-        "input_modalities": [
-          "text"
-        ],
-        "output_modalities": [
-          "text"
-        ],
-        "tokenizer": "Other",
-        "instruct_type": null
-      },
-      "pricing": {
-        "prompt": "0.0000000",
-        "completion": "0.0000000",
-        "request": "0",
-        "image": "0",
-        "web_search": "0",
-        "internal_reasoning": "0"
-      },
-      "top_provider": {
-        "context_length": 256000,
-        "max_completion_tokens": 32000,
-        "is_moderated": false
-      },
-      "per_request_limits": null,
-      "supported_parameters": [
-        "max_tokens",
-        "temperature",
-        "tools"
-      ],
-      "default_parameters": {},
-      "isFree": true
-    },
     {
       "id": "x-ai/grok-code-fast-1:optimized:free",
       "canonical_slug": "x-ai/grok-code-fast-1:optimized:free",