Arya is Gab AI’s frontier model.

An AI Image Generator.

Claude 3.7 Sonnet excels at coding and complex reasoning tasks, featuring a hybrid reasoning system that allows for both quick responses and extended step-by-step thinking, along with state-of-the-art performance in software engineering, a large 128K token context window, and improved instruction-following capabilities.

Claude 3.5 Haiku is Anthropic's fastest AI model, offering advanced coding, tool use, and reasoning capabilities.

ChatGPT-4o is OpenAI's newest flagship model that offers GPT-4-level intelligence with improved speed and enhanced capabilities across text and support for over 50 languages.

ChatGPT o1-mini is OpenAI's specialized reasoning model that excels at STEM tasks, particularly math and coding.

DeepSeek V3 is a powerful 671B parameter Mixture-of-Experts model that excels in mathematics and coding tasks, offering state-of-the-art performance. Updated March 24th. Hosted in 🇺🇸

DeepSeek R1 is an open-source large language model that excels at complex reasoning tasks, utilizing a novel reinforcement learning approach to achieve performance comparable to OpenAI's o1 across math, coding, and general knowledge benchmarks, while offering greater accessibility and customization potential for researchers and developers. Hosted in 🇺🇸

Perplexity AI is an advanced AI-powered search engine that provides conversational, real-time answers to queries by synthesizing information from multiple web sources, offering features like contextual follow-ups, and citation tracking.

Perplexity Sonar is a cutting-edge AI search technology that combines real-time web connectivity with advanced language models to provide fast, accurate, and cost-effective search capabilities, featuring customizable sources, citations, and a 127,000-token context window, while significantly outperforming similar models in speed and user satisfaction.

ChatGPT o3-mini is OpenAI's latest small reasoning model, offering advanced capabilities in STEM fields, coding, and logical problem-solving.

Gemini 2.0 Flash is Google's powerful workhorse model with low latency and enhanced performance, featuring a 1 million token context window.

Perplexity's R1 1776 is an open-source version of the DeepSeek-R1 language model that has been post-trained to remove Chinese Communist Party-aligned censorship while maintaining its powerful reasoning capabilities, offering unbiased and accurate responses on sensitive topics.

Grok 2 is xAI's advanced AI model featuring state-of-the-art capabilities in chat, coding, and reasoning, with real-time information access.

Gemini 1.5 Flash is Google's fast and efficient multimodal AI model, offering improved speed, a large 1 million token context window, and enhanced capabilities.

Qwen's QwQ-32B medium-sized reasoning model. Hosted in the US.

Google's Gemini 2.5 Pro Experimental (March 25th)