§ AI Wiki / Glossary
One-line definitions, the AI dictionary.
§ Search this category
Search the Wiki →OpenAI's chat product, released in November 2022, that brought AI assistants to mainstream scale.
The DeepMind result showing that for a fixed compute budget, parameters and training data should grow in balance.
The set of decisions that defines how a model's response is structured.
Anthropic's top-tier Claude model, built for the most complex reasoning and long-context tasks.
Anthropic's mid-to-upper tier Claude model balancing speed and capability, optimized for daily assistant and coding use.
Models capable of understanding or producing more than one input type — text, image, audio, video.
A technique that lets you condition diffusion models with structural inputs like pose, edges, or composition.
OpenAI's image-generation model series that brought text-to-image into public awareness.
An AI system designed and limited to perform a single task or narrow domain.
DeepSeek's 2025 open-weight reasoning model, which shook the industry with low-cost long-form thinking.
DeepSeek's late-2024 open-weight LLM, a strong and efficient Mixture-of-Experts model.
A machine learning paradigm where the model learns from labeled input-output pairs.
A machine learning paradigm in which the model finds patterns in data without any labels.
A subfield of machine learning that uses many-layered neural networks to automatically learn complex patterns.
An editing technique that extends an image beyond its original borders by generating new content.
A rating system from chess that derives relative skill scores from pairwise match outcomes.
A test suite that scores a model or system against predefined criteria.
A technique that embeds an invisible statistical signature into AI-generated text or images for later detection.
A 2024 image model from Black Forest Labs notable for photorealistic results.
A 3D synthesis method that represents scenes as clouds of 3D Gaussians for real-time rendering.
Google DeepMind's multimodal foundation model family and the assistant product built on top of it.
The frontier-tier flagship of the Gemini family, focused on multimodality, long context, and advanced reasoning.
A 2025 release of Z.AI's open-weight GLM language model family.
Google's AI research and product organization, behind Gemini, AlphaGo, and other landmark systems.