Models

	Provider	Model	Context	Capabilities
	OpenAI	GPT-5.2	400k	👁️🔧
	OpenAI	GPT-5.1	400k	👁️🔧
	Anthropic	Claude Sonnet 4.5	1M	👁️🔧
	Anthropic	Claude Haiku 4.5	200k	👁️🔧
	xAI	Grok 4.1 Fast	2M	👁️🔧
	xAI	Grok 4 Fast	2M	👁️🔧
	Google	Gemini 3 Pro	1M	👁️🔧
	Google	Gemini 3 Flash	1M	👁️🔧
	Meta	Llama 4 Scout 17B	328k	🔧
	Amazon	Nova 2 Lite	1M	👁️🔧
	Moonshot AI	Kimi K2.5	262k	🔧
	Qwen	Qwen3-2507	256k	🔧

And more…

​

Vision

Vision capabilities allow AI models to understand and analyze images. Models with vision support (indicated by 👁️ in the capabilities column) can:

• Analyze and describe images
• Extract text from images (OCR)
• Answer questions about visual content
• Understand charts, graphs, and diagrams
• Process screenshots and documents

Vision-enabled models are ideal for tasks like document analysis, image-based Q&A, and visual content understanding.

​

Tools

Tools (also known as function calling) enable AI models to interact with external systems and perform actions beyond text generation. Models with tool support (indicated by 🔧 in the capabilities column) can:

• Call external APIs and services
• Execute code and calculations
• Access real-time data
• Perform structured data operations
• Integrate with third-party applications

Tool-enabled models are essential for building AI agents, automation workflows, and applications that require dynamic interactions with external systems.

​

Web Search

Web search capabilities allow AI models to access real-time information from the internet, providing up-to-date answers and current data beyond their training knowledge. Key features:

• Real-time information: Access the latest news, events, and data
• Current facts: Get information that may have changed since the model’s training
• Comprehensive results: Search across multiple sources for accurate answers
• Citation support: Models can provide sources for their information

Web search is particularly useful for:

• Current events and news
• Real-time data (stock prices, weather, etc.)
• Recent developments in technology, science, and other fields
• Verifying information that may have changed

The model will automatically search the web when it needs current information to answer your questions accurately.

​

Context Window

The context window represents the maximum number of tokens (words or word pieces) a model can process in a single conversation. This includes both your input and the model’s output. Why it matters:

• Larger context windows (1M, 2M, 10M) allow you to work with longer documents, maintain longer conversations, and process more information at once
• Smaller context windows (128k, 200k) are more cost-effective and faster for shorter tasks

Choose a model with an appropriate context window based on your use case:

• Short tasks: 128k-200k is sufficient
• Document analysis: 400k-1M recommended
• Long conversations or large documents: 1M+ preferred

​

Model Parameters

Model parameters refer to the size and complexity of the neural network. While not always directly visible, parameter count affects:

• Model capability: Larger models generally have better reasoning and knowledge
• Response quality: More parameters often mean more nuanced and accurate responses
• Speed: Smaller models respond faster
• Cost: Larger models typically cost more per token

Common parameter sizes:

• Small (8B-17B): Fast, cost-effective, good for simple tasks
• Medium (20B-70B): Balanced performance and speed
• Large (70B+): Best quality, slower, higher cost

The parameter count is often reflected in the model name (e.g., “Llama 4 Scout 17B” has 17 billion parameters).