Help & Support
Comprehensive guides for using Insurgent Ink
Model Selection
Choose the right AI models for your translation needs, balancing quality, speed, and cost.
Model Selection Overview
Each AI provider offers multiple models with different capabilities, performance characteristics, and pricing. You can select a specific model for each API key, or use the provider's default model.
✨ New Feature
Insurgent Ink now supports per-API-key model selection, allowing you to customise your translation experience based on specific use cases and requirements.
Available Models by Provider
OpenAI Models
GPT-4o
RECOMMENDEDLatest and most capable model. Best overall balance of quality, speed, and features.
GPT-4o-mini
FAST & CHEAPFaster and more affordable version. Great for high-volume translations.
GPT-4-turbo
Previous generation flagship model. Still excellent for complex translations.
GPT-3.5-turbo
Budget option. Suitable for simple translations and high-volume work.
Anthropic Models
Claude-3-5-Sonnet
LATESTLatest and most sophisticated Claude model. Exceptional for nuanced and creative translations.
Claude-3-Haiku
FAST & EFFICIENTFastest and most cost-effective Claude model. Perfect for bulk translations and real-time processing.
Claude-3-Sonnet
Previous generation model. Still excellent balance of performance and cost.
Claude-3-Opus
Most capable Claude model for complex reasoning and analysis. Premium option for critical translations.
Other Providers
DeepL
SPECIALISTSingle dedicated translation model. Industry leader for supported language pairs.
Perplexity
RESEARCHMultiple models with real-time web search. Great for contextual and up-to-date translations.
OpenRouter
AGGREGATORAccess to dozens of models from different providers. Choose optimal model per task.
How to Select Models
Choose During API Key Setup
When adding a new API key, select your preferred model from the dropdown.
Leave empty to use the provider's default model
Model Shows in Translation
Your selected model will be displayed when choosing a provider for translation.
Switch Models Anytime
Update your model selection by editing your API key configuration.
Tip: You can have multiple API keys for the same provider with different models to easily switch between quality/speed/cost preferences.
Choosing the Right Model
Budget Conscious
Quality Priority
Speed Priority
Performance & Cost Considerations
Understanding the tradeoffs between different models helps you optimize for your specific needs:
💰 Cost Optimization
Costs vary by provider and may change. Check current pricing on provider websites.
⚡ Speed Comparison
Speed can vary based on content complexity and current API load.
Smart Model Selection Strategy
Use different models for different types of content to optimize quality and cost:
📄 Content Type Matching
- • Legal/Medical: GPT-4o, Claude-3-5-Sonnet for accuracy
- • Marketing: Claude models for creative flair
- • Technical docs: GPT-4o for precision
- • Bulk content: GPT-4o-mini, Claude-3-Haiku for efficiency
- • European languages: DeepL for specialized quality
🎯 Multi-Key Strategy
- • Development key: Fast, cheap model for testing
- • Production key: High-quality model for final work
- • Bulk key: Cost-effective model for volume work
- • Premium key: Best model for critical content
- • Backup key: Alternative provider for reliability
Custom and Self-Hosted Models
When using custom API endpoints, you can specify any model name that your API supports. This is perfect for organizations with specialized models or self-hosted solutions.
Self-Hosted Examples
- •
llama-2-70b-chat- Ollama deployment - •
mistral-7b-instruct- Mistral AI models - •
codellama-13b- Code-specialized models - •
local-translator-v2- Custom fine-tuned models
Enterprise Models
- •
company-legal-translator- Domain-specific - •
azure-gpt-4-deployment- Cloud hosted - •
custom-multilingual-v3- Internal development - •
compliance-translator- Regulated industries
💡 Custom Model Tips
- • Test thoroughly with your specific use case before production
- • Monitor performance and adjust timeout settings as needed
- • Consider fallback options for reliability
- • Document model capabilities and limitations for your team