Update LLM providers and constants (#1937)

- Updated constants in app/lib/.server/llm/constants.ts
- Modified stream-text functionality in app/lib/.server/llm/stream-text.ts
- Updated Anthropic provider in app/lib/modules/llm/providers/anthropic.ts
- Modified GitHub provider in app/lib/modules/llm/providers/github.ts
- Updated Google provider in app/lib/modules/llm/providers/google.ts
- Modified OpenAI provider in app/lib/modules/llm/providers/openai.ts
- Updated LLM types in app/lib/modules/llm/types.ts
- Modified API route in app/routes/api.llmcall.ts

app/lib/modules/llm/providers/anthropic.ts

@@ -22,6 +22,7 @@ export default class AnthropicProvider extends BaseProvider {
       label: 'Claude 3.5 Sonnet',
       provider: 'Anthropic',
       maxTokenAllowed: 200000,
+      maxCompletionTokens: 128000,
     },
 
     // Claude 3 Haiku: 200k context, fastest and most cost-effective
@@ -30,6 +31,7 @@ export default class AnthropicProvider extends BaseProvider {
       label: 'Claude 3 Haiku',
       provider: 'Anthropic',
       maxTokenAllowed: 200000,
+      maxCompletionTokens: 128000,
     },
   ];
 
@@ -84,6 +86,7 @@ export default class AnthropicProvider extends BaseProvider {
         label: `${m.display_name} (${Math.floor(contextWindow / 1000)}k context)`,
         provider: this.name,
         maxTokenAllowed: contextWindow,
+        maxCompletionTokens: 128000, // Claude models support up to 128k completion tokens
       };
     });
   }

app/lib/modules/llm/providers/github.ts

@@ -14,13 +14,31 @@ export default class GithubProvider extends BaseProvider {
 
   // find more in https://github.com/marketplace?type=models
   staticModels: ModelInfo[] = [
-    { name: 'gpt-4o', label: 'GPT-4o', provider: 'Github', maxTokenAllowed: 8000 },
-    { name: 'o1', label: 'o1-preview', provider: 'Github', maxTokenAllowed: 100000 },
-    { name: 'o1-mini', label: 'o1-mini', provider: 'Github', maxTokenAllowed: 8000 },
-    { name: 'gpt-4o-mini', label: 'GPT-4o Mini', provider: 'Github', maxTokenAllowed: 8000 },
-    { name: 'gpt-4-turbo', label: 'GPT-4 Turbo', provider: 'Github', maxTokenAllowed: 8000 },
-    { name: 'gpt-4', label: 'GPT-4', provider: 'Github', maxTokenAllowed: 8000 },
-    { name: 'gpt-3.5-turbo', label: 'GPT-3.5 Turbo', provider: 'Github', maxTokenAllowed: 8000 },
+    { name: 'gpt-4o', label: 'GPT-4o', provider: 'Github', maxTokenAllowed: 128000, maxCompletionTokens: 16384 },
+    { name: 'o1', label: 'o1-preview', provider: 'Github', maxTokenAllowed: 100000, maxCompletionTokens: 16384 },
+    { name: 'o1-mini', label: 'o1-mini', provider: 'Github', maxTokenAllowed: 65536, maxCompletionTokens: 8192 },
+    {
+      name: 'gpt-4o-mini',
+      label: 'GPT-4o Mini',
+      provider: 'Github',
+      maxTokenAllowed: 128000,
+      maxCompletionTokens: 16384,
+    },
+    {
+      name: 'gpt-4-turbo',
+      label: 'GPT-4 Turbo',
+      provider: 'Github',
+      maxTokenAllowed: 128000,
+      maxCompletionTokens: 8192,
+    },
+    { name: 'gpt-4', label: 'GPT-4', provider: 'Github', maxTokenAllowed: 8192, maxCompletionTokens: 8192 },
+    {
+      name: 'gpt-3.5-turbo',
+      label: 'GPT-3.5 Turbo',
+      provider: 'Github',
+      maxTokenAllowed: 16385,
+      maxCompletionTokens: 4096,
+    },
   ];
 
   getModelInstance(options: {

app/lib/modules/llm/providers/google.ts

@@ -17,10 +17,22 @@ export default class GoogleProvider extends BaseProvider {
      * Essential fallback models - only the most reliable/stable ones
      * Gemini 1.5 Pro: 2M context, excellent for complex reasoning and large codebases
      */
-    { name: 'gemini-1.5-pro', label: 'Gemini 1.5 Pro', provider: 'Google', maxTokenAllowed: 2000000 },
+    {
+      name: 'gemini-1.5-pro',
+      label: 'Gemini 1.5 Pro',
+      provider: 'Google',
+      maxTokenAllowed: 2000000,
+      maxCompletionTokens: 32768,
+    },
 
     // Gemini 1.5 Flash: 1M context, fast and cost-effective
-    { name: 'gemini-1.5-flash', label: 'Gemini 1.5 Flash', provider: 'Google', maxTokenAllowed: 1000000 },
+    {
+      name: 'gemini-1.5-flash',
+      label: 'Gemini 1.5 Flash',
+      provider: 'Google',
+      maxTokenAllowed: 1000000,
+      maxCompletionTokens: 32768,
+    },
   ];
 
   async getDynamicModels(
@@ -89,11 +101,19 @@ export default class GoogleProvider extends BaseProvider {
       const maxAllowed = 2000000; // 2M tokens max
       const finalContext = Math.min(contextWindow, maxAllowed);
 
+      // Get completion token limit from Google API
+      let completionTokens = 32768; // default fallback
+
+      if (m.outputTokenLimit && m.outputTokenLimit > 0) {
+        completionTokens = Math.min(m.outputTokenLimit, 128000); // Cap at reasonable limit
+      }
+
       return {
         name: modelName,
         label: `${m.displayName} (${finalContext >= 1000000 ? Math.floor(finalContext / 1000000) + 'M' : Math.floor(finalContext / 1000) + 'k'} context)`,
         provider: this.name,
         maxTokenAllowed: finalContext,
+        maxCompletionTokens: completionTokens,
       };
     });
   }

app/lib/modules/llm/providers/openai.ts

@@ -17,10 +17,16 @@ export default class OpenAIProvider extends BaseProvider {
      * Essential fallback models - only the most stable/reliable ones
      * GPT-4o: 128k context, high performance, recommended for most tasks
      */
-    { name: 'gpt-4o', label: 'GPT-4o', provider: 'OpenAI', maxTokenAllowed: 128000 },
+    { name: 'gpt-4o', label: 'GPT-4o', provider: 'OpenAI', maxTokenAllowed: 128000, maxCompletionTokens: 16384 },
 
     // GPT-3.5-turbo: 16k context, fast and cost-effective
-    { name: 'gpt-3.5-turbo', label: 'GPT-3.5 Turbo', provider: 'OpenAI', maxTokenAllowed: 16000 },
+    {
+      name: 'gpt-3.5-turbo',
+      label: 'GPT-3.5 Turbo',
+      provider: 'OpenAI',
+      maxTokenAllowed: 16000,
+      maxCompletionTokens: 4096,
+    },
   ];
 
   async getDynamicModels(