AI Model
Reference
Complete guide to all 40+ AI models available in FRENZY.BOT. Each model is hand-picked for production readiness and performance.
Anthropic (Claude)
Advanced reasoning models with superior analytical capabilities and safety features.
Claude Opus 4.6
Complex reasoning, advanced software engineering, agentic workflows, long-horizon computer use
Superior analytical capabilities, excels at multi-step problem solving, ideal for complex technical tasks and research
Claude Sonnet 4.6
Balanced performance for business applications, content creation, complex dialogue
Latest generation with improved reasoning and instruction following, excellent for customer support and content generation
Claude Sonnet 4
General business tasks, customer support, content creation
Reliable performance, excellent instruction following, cost-effective for production use
Claude Sonnet 4.5
Business applications, analytical tasks, complex Q&A
Strong reasoning capabilities, good balance of performance and cost
Claude Haiku 4.5
Fast responses, real-time chat, high-volume interactions
Sub-second response times, cost-effective for scale, good for simple queries
Claude 3.5 Haiku
Budget-conscious real-time applications, quick responses
Excellent speed-to-cost ratio, reliable for simple tasks
OpenAI (GPT)
Industry-leading models with exceptional performance across diverse tasks and applications.
GPT-5.2
Advanced reasoning, complex tasks, professional applications
Latest generation with superior capabilities, excellent for complex problem solving
GPT-5.1
Advanced applications, complex reasoning, professional use
Strong performance across diverse tasks, good for business-critical applications
GPT-5
General advanced tasks, business applications, content creation
Excellent all-around performance, reliable for production use
GPT-4.1
Advanced instruction following, software engineering, long-context reasoning
Flagship model with excellent reliability, strong for technical and analytical tasks
o3
Complex reasoning, scientific tasks, mathematical problem solving
Advanced reasoning capabilities, excellent for complex analytical work
o4 Mini
Balanced reasoning and speed, complex tasks with cost control
Good reasoning capabilities with better cost efficiency than full o-series models
GPT-5 Mini
Cost-effective advanced tasks, business applications
Excellent value, good capabilities for the price point
GPT-4.1 Mini
Balanced performance and cost, general business tasks
Good cost-to-performance ratio, reliable for production use
GPT-5 Nano
High-volume simple tasks, budget operations
Extremely cost-effective, good for basic Q&A and simple interactions
GPT-4.1 Nano
Budget operations, simple queries, high-volume applications
Very cost-effective with good context window, ideal for scale
Google (Gemini)
Multimodal models with exceptional reasoning and integration with Google ecosystem.
Gemini 3 Pro Preview
Advanced reasoning, coding, mathematics, scientific tasks
State-of-the-art reasoning capabilities, excellent for complex technical and analytical work
Gemini 2.5 Pro
Advanced reasoning, coding, mathematics, scientific analysis
Strong performance across technical domains, good balance of capability and cost
Gemini 3 Flash Preview
Real-time applications, fast responses, high-volume interactions
Excellent speed with good reasoning capabilities, ideal for customer-facing chat
Gemini 2.5 Flash
Fast responses, general tasks, cost-effective operations
Good balance of speed and capability, cost-effective for scale
Gemini 2.5 Flash Lite
High-volume simple queries, budget operations
Extremely cost-effective, good for basic Q&A and simple tasks
DeepSeek
Open-source models with exceptional reasoning capabilities and cost efficiency.
DeepSeek R1
Complex reasoning, mathematical problems, coding challenges
Open-source reasoning model with performance on par with OpenAI o1, excels at logical deduction and step-by-step problem solving
DeepSeek V3.2
General AI tasks, content generation, customer support
Excellent value proposition, strong performance across diverse tasks, improved long-context handling
DeepSeek V3.1
Budget-conscious applications, standard chatbot tasks
Outstanding cost efficiency, good performance for routine tasks
Meta (Llama)
Open-weights models with transparency and robust performance for diverse applications.
Llama 4 Maverick
General AI tasks, content creation, customer support
Open-weights transparency, robust performance, excellent value
Llama 4 Scout
Lightweight applications, fast responses, budget operations
Very cost-effective, good for standard chatbot tasks
Qwen (Alibaba)
Advanced models with exceptional reasoning and multilingual capabilities.
Qwen3 Max
Advanced reasoning, multilingual tasks, complex problem solving
Major improvements in reasoning, instruction following, multilingual support, excellent for math, coding, and science
Qwen3.5 397B A17B
General advanced tasks, business applications, content creation
Strong performance with good cost efficiency, reliable for diverse use cases
Qwen3.5 Plus 2026-02-15
Latest performance, general tasks, large document processing
Recent release with improved capabilities, excellent context window
Qwen3 Coder 480B
Coding tasks, technical applications, software development
Mixture-of-Experts design optimized for agentic coding tasks, function calling, and tool use
Qwen3 Next 80B A3B Instruct
Cost-effective general tasks, lightweight applications
Sparse MoE design with efficient inference, good performance for the cost
xAI (Grok)
Real-time knowledge models with massive context windows and fast processing.
Grok 4.1 Fast
Real-time applications, large document processing, fast responses
Massive context window, excellent speed, good for high-volume interactions
Grok 4 Fast
Fast responses, large context applications, real-time processing
Excellent speed with massive context, ideal for document analysis
Grok 3 Mini
Cost-effective real-time knowledge, fast responses
Good balance of speed and cost, real-time knowledge access
MiniMax
Versatile models with strong performance across multiple languages and tasks.
MiniMax-01
General tasks, content generation, multilingual applications
Strong overall performance, competitive pricing, good for diverse use cases
MiniMax M2.5
Coding tasks, technical applications, complex reasoning
Scoring 80.2% on SWE-Bench Verified, excellent for coding and technical problem solving
MiniMax M2.1
General applications, balanced performance
Good balance of capabilities and cost, reliable for diverse tasks
MiniMax M2-her
Conversational AI, dialogue-first applications
Optimized for immersive dialogue, excellent for chatbot interactions
Mistral
European models with strong multilingual support and reliable performance.
Mistral Medium 3.1
General business tasks, content creation, European languages
Strong multilingual support, reliable performance, good for European markets
Mistral Medium 3
Business applications, content generation
Consistent performance, good for professional use cases
Mistral Small 3.2 24B
Lightweight applications, fast responses, budget operations
Very cost-effective, good for simple queries and high-volume use
Z.ai
Reliable models with good performance across general AI tasks.
GLM 5
General AI tasks, multilingual applications, content generation
Strong overall performance, good for diverse use cases
GLM 4.7
Balanced performance, business applications, technical tasks
Reliable performance with good reasoning capabilities
GLM 4.7 Flash
Fast responses, budget operations, high-volume interactions
Very cost-effective, good for simple queries and scale
Model Selection Guide
By Use Case
Customer Support & Chat
Claude Sonnet 4.6, GPT-4.1, Gemini 2.5 Flash
Fast response times, Reliable performance, Cost-effective at scale
Complex Reasoning & Analysis
Claude Opus 4.6, GPT-5.2, Gemini 3 Pro, DeepSeek R1
Superior analytical capabilities, Multi-step problem solving, High accuracy on difficult tasks
Content Creation & Writing
Claude Sonnet 4, GPT-5, Gemini 2.5 Pro
Natural language generation, Creative capabilities, Consistent quality
Coding & Technical Tasks
Qwen3 Coder, MiniMax M2.5, GPT-5.2, DeepSeek R1
Strong code understanding, Tool use and function calling, Technical problem solving
Budget-Conscious Operations
DeepSeek V3.1, GPT-5 Nano, Gemini 2.5 Flash Lite, Mistral Small 3.2
Excellent cost efficiency, Good performance for routine tasks, Scalable for high volume
Large Document Processing
Grok 4.1 Fast, Claude Opus 4.6, Gemini 2.5 Pro
Massive context windows, Long-document understanding, Comprehensive analysis
By Performance Tier
Premium Tier (Highest Performance)
PREMIUMClaude Opus 4.6, GPT-5.2, Gemini 3 Pro, Claude Sonnet 4.6
Standard Tier (Balanced Performance)
STANDARDGPT-5, GPT-4.1, Gemini 2.5 Pro, Claude Sonnet 4, Qwen3 Max
Value Tier (Best Cost/Performance)
VALUEDeepSeek V3.2, GPT-5 Mini, Gemini 2.5 Flash, Claude 3.5 Haiku, Llama 4 Maverick
Budget Tier (Lowest Cost)
BUDGETDeepSeek V3.1, GPT-5 Nano, Gemini 2.5 Flash Lite, Mistral Small 3.2, GLM 4.7 Flash
Pricing Calculator
Use this quick reference to estimate costs:
Input Costs (per 1M tokens)
Output Costs (per 1M tokens)
Technical Specifications
Context Windows
Zero Data Retention Support
Models marked with ✅ support ZDR through OpenRouter's provider.zdr: true parameter, ensuring your data is never stored or used for training.
Zero Data Retention available on select models for maximum privacy.
Frequently Asked Questions
How often are new models added?
New models are evaluated and added within 24 hours of stable release from major providers.
Can I switch between models?
Yes, you can change models at any time in Settings → AI Engine. Changes take effect for new conversations.
Which models support Zero Data Retention?
Models marked with ✅ ZDR support include most Claude, GPT, DeepSeek, and select models from other providers.
How do I choose the right model for my use case?
Consider your budget, required response speed, complexity of tasks, and volume. Use the selection guide above or start with Standard tier models.
Are there usage limits?
Usage is billed per token through OpenRouter. FRENZY.BOT adds no markup. Check OpenRouter for current pricing and any rate limits.
Can I use multiple models for different tasks?
Currently, one model is selected per bot. Choose the model that best fits your primary use case.
Ready to Choose Your AI Model?
Get started with FRENZY.BOT and deploy any of these powerful AI models for your business.
Last updated: February 2026
Model availability and pricing are subject to change. Check OpenRouter for current information.