What is AI Chat?

The AI Chat tool allows you to interact with advanced language models to ask questions, draft content, analyze data, or brainstorm ideas. You can type requests, attach supporting files, and receive contextual responses in real time.

Input Types

Text-to-Text

• Enter queries in plain language.

• Supports both short prompts and multi-paragraph requests.

Attachments

• Supported file types: .jpg, .jpeg, .png, .gif, .webp, .pdf, .txt, .csv, .json, .md, .html, .docx, .pptx, .xlsx, .epub

• Supported image extensions

• File size: up to 20 MB per file.

Chatly supports a wide range of models from leading providers. Each model has different strengths in speed, reasoning, creativity, and efficiency. This section outlines the available models and explains their intended use.

OpenAI Models

• GPT-5 Mini: A smaller unified model suited for everyday tasks where quick responses are more important than deep reasoning.

• GPT-5: OpenAI’s unified model. Provides advanced reasoning, long-form answers, and versatility across domains.

• GPT-5.1: An upgraded flagship model that blends faster responsiveness with deeper adaptive reasoning. More efficient, more conversational, and better at complex multi-step tasks than GPT-5.

• GPT-o3: A reasoning-focused model, designed for solving logic-heavy problems and structured analysis.

• GPT-4o-mini: Lightweight and fast. Best used for short queries and simple outputs.

• GPT-4.1-nano: The smallest GPT-4.1 release. Prioritizes efficiency while maintaining acceptable accuracy.

• GPT-4.1-mini: A compact version of GPT-4.1. Provides a balance between output quality and processing speed.

• GPT-4.1: OpenAI’s most advanced GPT-4.1 model. Offers high accuracy and detailed results.

• GPT-4o: Multimodal model capable of handling text and other input formats. Provides advanced reasoning and content generation.

• GPT-5.2: OpenAI’s next-generation flagship model, offering improved reasoning, long-context understanding, advanced coding, and reliable tool use for professional and everyday tasks.

• GPT-5.2 Pro: OpenAI's most advanced and most reliable GPT-5.2 model, delivering the highest accuracy and strongest performance on scientific, mathematical, and reasoning benchmarks.

Anthropic Models

• Claude 4.5 Sonnet: Powerful model for coding, reasoning, and complex problem solving. It delivers state-of-the-art performance across real-world tasks and domain knowledge in finance, law, medicine, and STEM.

• Claude 4 Sonnet: Anthropic’s advanced model. Strong performance in reasoning, analysis, and programming tasks.

• Claude 4.5 Haiku: Anthropic’s fastest, cost-efficient model.

• Claude 3.7 Sonnet: A stable, well-rounded model offering balanced performance for general use.

• Claude 3.5 Haiku: Anthropic’s fast, cost-efficient model.

• Claude 4.5 Opus: Anthropic's most capable AI model that excels at coding, agent tasks, and everyday work like research and document handling.

Google Models

• Gemini 2.5 Pro: Supports advanced reasoning, detailed answers, and multimodal input.

• Gemini 3 Pro: Google's most advanced AI model, optimized for high-level reasoning, complex tasks, and expert-level knowledge across various domains.

• Gemini 3.0 Flash: Google’s ultra-fast multimodal AI model, optimized for low-latency responses, real-time interactions, and high-throughput tasks at scale.

xAI Models

• Grok 4: The latest Grok release. Optimized for advanced reasoning and problem-solving.

• Grok 4 Fast: Delivers quick, reliable answers to help you stay informed, research faster, and solve problems efficiently across a wide range of topics.

• Grok 3 Mini: A smaller Grok variant. Provides faster responses while retaining reasoning capabilities.

• Grok 4.1 Fast (Reasoning): X.AI's most advanced artificial intelligence model to date, designed for high-level reasoning, real-time information access, and complex problem-solving.

Chatly Model

• Chatly-o1: An open-source model developed by Chatly. Optimized for reasoning while offering transparency and flexibility.

MoonShot AI Model

• Kimi K2 (Instruct): A 1-trillion-parameter, Mixture-of-Experts model from Moonshot AI (32B active), tuned for fast, instruction-following conversation, tool use, and agentic workflows.