LLM Providers Reference
Complete reference for all 9LLM providers supported by ScreenerBot. Learn about recommended models and setup instructions for each provider. Check each provider's own website for current pricing, free tiers, and rate limits, since these change over time.
Provider Comparison
| Provider | Models | Best For |
|---|---|---|
OpenRouterRecommended | Unified access to most other providers | One key, easy model switching |
Anthropic | Claude models | Best reasoning |
OpenAI | GPT models | Highest accuracy |
Ollama | Any local model you choose | Local/private, self-hosted |
Gemini | Google's Gemini models | Large context analysis |
Groq | Llama, Mixtral, Gemma | Low-latency inference |
DeepSeek | DeepSeek's own models | High-volume screening |
Mistral | Mistral's Large/Small models | EU-based provider |
Together AI | Open-source models | Wide model selection |
Provider Setup Guides
OpenRouter (Recommended)
Details:
- • Models: Access to Anthropic, OpenAI, Google, Meta and other brands via one API
- • Website: openrouter.ai
Model Selection:
Route to almost any model on the market through a single key. Browse OpenRouter's model list for current options.
Why choose OpenRouter: One key gives you access to nearly every major model, so you can switch providers or models without reconfiguring anything. This is the recommended default for most ScreenerBot users.
Anthropic (Claude)
Details:
- • Models: Multiple Claude tiers at different cost/quality points
- • Website: console.anthropic.com
Model Selection:
Claude tiers range from fast & cheap to most capable. Check Anthropic's model list for current options.
Why choose Anthropic: Superior reasoning abilities for complex risk assessment. Claude excels at identifying subtle scam indicators and providing nuanced analysis.
OpenAI
Details:
- • Models: Multiple GPT tiers at different cost/quality points
- • Website: platform.openai.com
Model Selection:
OpenAI offers tiers from cheapest to most capable. Check OpenAI's model list for current options.
Why choose OpenAI: Industry-leading accuracy for complex analysis, with a model tier to fit most budgets.
Ollama (Local)
Details:
- • Privacy: 100% local, no API calls
- • Requirements: 8GB+ RAM recommended
- • Website: ollama.com
Model Selection:
Pull any open-source model that fits your hardware. Browse Ollama's model library for current options.
Why choose Ollama: Perfect for privacy-conscious traders or those who want to run everything on their own hardware, with no external API calls.
Installation:
# Install Ollama
curl -fsSL https://ollama.com/install.sh | sh
# Pull a model (see ollama.com/library for current options)
ollama pull <model-name>
# Verify it's running
ollama listGoogle Gemini
Details:
- • Context: Up to 2M tokens on some tiers
- • Models: Multiple Gemini tiers
- • Website: aistudio.google.com
Model Selection:
Balanced and most-capable tiers are both available. Check Google's Gemini model list for current options.
Why choose Gemini: Massive context window allows analyzing tokens with extensive historical data.
Groq
Details:
- • Speed: Very low latency inference
- • Models: Llama, Mixtral, Gemma
- • Website: console.groq.com
Model Selection:
Several open-source families (Llama, Mixtral, Gemma) at different size/speed tradeoffs. Check Groq's model list for current options.
Why choose Groq: Useful when latency matters more than model quality.
DeepSeek
Details:
- • Models: Chat and reasoning-focused variants
- • Website: platform.deepseek.com
Model Selection:
A general chat model plus an enhanced-reasoning variant. Check DeepSeek's docs for current model IDs.
Why choose DeepSeek: A cost-conscious option for high-volume screening.
Mistral AI
Details:
- • Region: European provider
- • Models: Large, Medium and Small tiers
- • Website: console.mistral.ai
Model Selection:
Pick the Large tier for quality or Small for speed and cost. Check Mistral's model list for current options.
Why choose Mistral: A European alternative with fast inference and multilingual capabilities.
Together AI
Details:
- • Models: Open-source models
- • Website: api.together.xyz
Model Selection:
Llama, Mixtral, Qwen and other open-source families are all available. Check Together AI's model list for current options.
Why choose Together AI: A wide selection of open-source models under one account.
Choosing the Right Provider
Recommended Default
If you're not sure which to pick:
- • OpenRouter - one key, access to almost every model
For Best Reasoning
Best options for complex analysis and risk assessment:
- • Anthropic Claude - best reasoning
- • OpenAI - highest accuracy
For Local / Private
Run everything on your own hardware:
- • Ollama - fully local, no external API calls
For Large Context
Analyzing tokens with extensive historical data:
- • Gemini - very large context windows