chore: Update default LLM model from mistral-small:24b to qwen3:14b

Author: hydropix

.env.example

@@ -1,6 +1,6 @@
 # Translation API Configuration
 API_ENDPOINT=http://localhost:11434/api/generate
-DEFAULT_MODEL=mistral-small:24b
+DEFAULT_MODEL=qwen3:14b
 
 # Server Configuration
 PORT=5000  # Port for the web interface

DOCKER.md

@@ -18,7 +18,7 @@ docker run -d \
   -v $(pwd)/translated_files:/app/translated_files \
   -v $(pwd)/logs:/app/logs \
   -e API_ENDPOINT=http://host.docker.internal:11434/api/generate \
-  -e DEFAULT_MODEL=mistral-small:24b \
+  -e DEFAULT_MODEL=qwen3:14b \
   ghcr.io/bropro/translatebookwithllm:latest
 ```
 
@@ -44,7 +44,7 @@ Edit the `.env` file to configure your LLM settings:
 
 ```env
 API_ENDPOINT=http://host.docker.internal:11434/api/generate
-DEFAULT_MODEL=mistral-small:24b
+DEFAULT_MODEL=qwen3:14b
 LLM_PROVIDER=ollama
 PORT=5000
 OLLAMA_NUM_CTX=2048
@@ -95,7 +95,7 @@ Docker automatically pulls the correct architecture for your system.
 | Variable | Description | Default |
 |----------|-------------|---------|
 | `API_ENDPOINT` | LLM API endpoint | `http://localhost:11434/api/generate` |
-| `DEFAULT_MODEL` | Default LLM model | `mistral-small:24b` |
+| `DEFAULT_MODEL` | Default LLM model | `qwen3:14b` |
 | `LLM_PROVIDER` | Provider (ollama/gemini/openai) | `ollama` |
 | `GEMINI_API_KEY` | Gemini API key | - |
 | `OPENAI_API_KEY` | OpenAI API key | - |
@@ -132,7 +132,7 @@ chmod 755 translated_files logs data
 docker run -d \
   -p 5000:5000 \
   -e API_ENDPOINT=http://host.docker.internal:11434/api/generate \
-  -e DEFAULT_MODEL=mistral-small:24b \
+  -e DEFAULT_MODEL=qwen3:14b \
   ghcr.io/bropro/translatebookwithllm:latest
 ```
 
@@ -157,7 +157,7 @@ services:
       - "5000:5000"
     environment:
       - API_ENDPOINT=http://ollama:11434/api/generate
-      - DEFAULT_MODEL=mistral-small:24b
+      - DEFAULT_MODEL=qwen3:14b
     depends_on:
       - ollama
 

deployment/.env.docker.example

@@ -22,7 +22,7 @@ LLM_PROVIDER=ollama
 API_ENDPOINT=http://host.docker.internal:11434/api/generate
 
 # Model to use (must be installed in your local Ollama)
-DEFAULT_MODEL=mistral-small:24b
+DEFAULT_MODEL=qwen3:14b
 
 # Context window size (8192 recommended for chunk_size=25)
 OLLAMA_NUM_CTX=8192

deployment/TESTING.md

@@ -252,7 +252,7 @@ docker-compose logs
    # Linux:
    # API_ENDPOINT=http://172.17.0.1:11434/api/generate
 
-   DEFAULT_MODEL=mistral-small:24b
+   DEFAULT_MODEL=qwen3:14b
    ```
 
 3. Verify Ollama is accessible from container:

deployment/docker-compose.yml

@@ -23,7 +23,7 @@ services:
       # Use host.docker.internal to access host machine from Docker on Windows/Mac
       # On Linux, use --network host or the host's IP address
       - API_ENDPOINT=${API_ENDPOINT:-http://host.docker.internal:11434/api/generate}
-      - DEFAULT_MODEL=${DEFAULT_MODEL:-mistral-small:24b}
+      - DEFAULT_MODEL=${DEFAULT_MODEL:-qwen3:14b}
       - OLLAMA_NUM_CTX=${OLLAMA_NUM_CTX:-8192}
 
       # Gemini Configuration (if using Gemini provider)

src/config.py

@@ -34,7 +34,7 @@
     print("🔧 DEFAULT SETTINGS BEING USED:")
     print(f"   • API Endpoint: http://localhost:11434/api/generate")
     print(f"   • LLM Provider: ollama")
-    print(f"   • Model: mistral-small:24b")
+    print(f"   • Model: qwen3:14b")
     print(f"   • Port: 5000")
     print(f"\n💡 TIP: If using a remote server or different provider, you MUST")
     print(f"   create a .env file with the correct settings.\n")
@@ -55,7 +55,7 @@
 
 # Load from environment variables with defaults
 API_ENDPOINT = os.getenv('API_ENDPOINT', 'http://localhost:11434/api/generate')
-DEFAULT_MODEL = os.getenv('DEFAULT_MODEL', 'mistral-small:24b')
+DEFAULT_MODEL = os.getenv('DEFAULT_MODEL', 'qwen3:14b')
 PORT = int(os.getenv('PORT', '5000'))
 MAIN_LINES_PER_CHUNK = int(os.getenv('MAIN_LINES_PER_CHUNK', '25'))
 REQUEST_TIMEOUT = int(os.getenv('REQUEST_TIMEOUT', '900'))

src/core/context_optimizer.py

@@ -158,7 +158,7 @@ def get_max_model_context(model_name: str) -> int:
     Get maximum context size for a model based on its family.
 
     Args:
-        model_name: Name of the model (e.g., "mistral-small:24b")
+        model_name: Name of the model (e.g., "qwen3:14b")
 
     Returns:
         Maximum context size in tokens

src/utils/env_helper.py

@@ -61,7 +61,7 @@ def validate_env_config(verbose: bool = True) -> dict:
     # Check critical configuration
     api_endpoint = os.getenv('API_ENDPOINT', 'http://localhost:11434/api/generate')
     llm_provider = os.getenv('LLM_PROVIDER', 'ollama')
-    default_model = os.getenv('DEFAULT_MODEL', 'mistral-small:24b')
+    default_model = os.getenv('DEFAULT_MODEL', 'qwen3:14b')
     gemini_key = os.getenv('GEMINI_API_KEY', '')
     openai_key = os.getenv('OPENAI_API_KEY', '')
 
@@ -142,7 +142,7 @@ def interactive_env_setup():
 
     if config['LLM_PROVIDER'] == 'ollama':
         config['API_ENDPOINT'] = input("   Ollama API endpoint [http://localhost:11434/api/generate]: ").strip() or 'http://localhost:11434/api/generate'
-        config['DEFAULT_MODEL'] = input("   Model name [mistral-small:24b]: ").strip() or 'mistral-small:24b'
+        config['DEFAULT_MODEL'] = input("   Model name [qwen3:14b]: ").strip() or 'qwen3:14b'
 
     elif config['LLM_PROVIDER'] == 'gemini':
         config['GEMINI_API_KEY'] = input("   Gemini API Key: ").strip()

src/web/static/js/providers/model-detector.js

@@ -9,7 +9,7 @@ import { DomHelpers } from '../ui/dom-helpers.js';
 
 /**
  * Extract parameter size from model name
- * @param {string} modelName - Model name (e.g., "mistral-small:7b", "llama-12b")
+ * @param {string} modelName - Model name (e.g., "qwen3:14b", "llama-12b")
  * @returns {number|null} Size in billions of parameters, or null if not detected
  */
 function extractModelSize(modelName) {