AI Tools Directory

An AI startup founded by Fei-Fei Li that builds world models capable of generating and reasoning about immersive, editable 3D environments. Its first product lets users create downloadable 3D scenes from prompts.

Autodesk's generative AI model trained on geometric data that can reason about components and systems to generate functional 3D models with an understanding of real-world design constraints.

An AI agent that can take a 2D image and autonomously reconstruct it as an editable 3D scene in Blender, enabling interactive physics simulations and scene manipulation.

A 3D rendering technique that builds scenes from millions of tiny elliptical 3D bumps (Gaussians) to create highly realistic virtual humans with subsurface scattering for skin and realistic hair rendering.

A Google DeepMind model announced in August that can generate dynamic, playable worlds from text prompts or images, retaining consistency for a few minutes at 720p resolution.

An AI aggregator platform that provides access to 25+ leading AI models including GPT-4o, Claude, Gemini, DeepSeek, Llama, and Perplexity in a single dashboard, allowing users to compare outputs side by side. It includes prompt engineering tools, image and document chat, web search, saved chat history, and a Chrome extension.

An AI aggregation platform that provides access to multiple AI models including ChatGPT, Gemini, and Mistral in one central hub, enabling text generation, image creation, video production, and SEO keyword research without needing separate subscriptions.

Japan's most comprehensive LLM evaluation platform, assessing language models across approximately 40 benchmarks covering agent capabilities (code generation, math reasoning, tool use) and alignment (instruction following, bias, toxicity, truthfulness, robustness).

A Google-owned platform for data science and machine learning that hosts competitions, datasets, and model repositories, enabling developers to download and deploy AI models.

A directory and news website that curates and catalogs AI tools, providing a searchable database alongside AI-related news coverage.

A platform for hosting, sharing, and discovering AI models, datasets, and machine learning applications, serving as a major hub for the open-source AI community.

An API aggregator platform that provides unified access to a wide range of AI models from multiple providers through a single interface.

A platform that tracks monthly unique visitors and usage statistics for various AI tools and platforms.

An AI benchmark platform that works like a blind taste test of models, where users compare two anonymous responses and vote on which is better, with leaderboards determined by user preference.

An AI-powered productivity and collaboration workspace that combines document creation, note organization, task planning, and visualization boards with built-in AI assistance for summarizing, rewriting, and refining ideas. It offers unlimited AI requests, real-time multiplayer editing, and shared workspaces.

An AI-powered legal assistant built on large language models that helps legal professionals with research, drafting, and analysis of legal documents.

A free online chat platform by Alibaba that allows users to interact with Qwen models including Qwen 3.5 without needing to download or run them locally.

An open-source C/C++ inference engine originally created by Georgi Gerganov (GGML) that enables efficient local inference of large language models on consumer hardware. It is widely regarded as the fundamental building block for running AI models locally.

A tensor library written in C for machine learning, created by Georgi Gerganov, that serves as the underlying framework for llama.cpp and enables efficient on-device AI inference. Its GGUF format has become a widely adopted standard for quantized model distribution.

A PyTorch-based on-device inference framework developed by Meta that enables running AI models efficiently on edge devices. It has adopted GGML's GGUF format as a preferred default for on-device inference.

An AI-powered financial reporting platform that helps companies and accounting firms automate manual aspects of financial statement preparation, including math verification and formatting. It raised $14.5 million in Series A funding led by Norwest.

A five-week Google-run cohort program that gives independent filmmakers access to Google's suite of AI tools — including Gemini, Nano Banana Pro, and Veo — to produce short films.

A U.S.-based AI chipmaker that designs and manufactures specialized hardware for AI computing, including wafer-scale processors optimized for training and inference of large-scale AI models.

A privacy-focused AI desktop assistant that runs entirely on the user's local computer without cloud connectivity, supporting chat, writing, coding, and document analysis offline with no subscription or token limits.

A cloud-based AI coding agent platform by Warp that lets users spin up multiple autonomous agents in isolated Docker containers, schedule them to run on cadences, and manage them from a central panel or via Slack and GitHub triggers.

An open-source library by NVIDIA for designing and generating synthetic training data from scratch or from seed data, part of the NeMo microservices ecosystem.

An open-source library that provides approximately 2x faster LLM training and 60% less VRAM usage compared to standard fine-tuning methods, enabling cost-effective model training.

A fully managed cloud GPU service by Hugging Face that allows users to submit and run AI model training jobs on cloud infrastructure with monitoring capabilities.

A London-based startup building an on-device AI inference engine optimized for Apple Silicon, enabling developers to run AI models efficiently on phones and laptops with a simple SDK integration. Built in Rust, it claims up to 37% faster model generation speeds.

An AI workforce management platform that lets organizations manage, coordinate, and oversee AI agents across teams and departments, serving as a system of record for AI employees regardless of which tools built them.

An AI product by Reload that acts as a software architect agent, maintaining shared project-level context and system requirements across coding agents and sessions to keep development consistent. It integrates as an extension in AI code editors like Cursor and Windsurf.

An AI-powered search feature by Reddit that surfaces community-recommended products with interactive carousels, pricing, images, and direct purchase links based on discussions across the platform.

Smart glasses by Indian AI company Sarvam designed to bring on-device AI models directly to users, built and designed in India.

An AI-powered meeting notetaker and intelligence platform that automatically captures meeting notes, provides contextual insights, and includes a sales tool that predicts deal outcomes using CRM data from HubSpot and Salesforce.

An AI-powered customer support tooling platform that automates customer service tasks including voice AI communication, enabling support agents to take on higher-value roles like supervision and relationship building.

NVIDIA's global startup support program that provides AI startups with technical expertise, go-to-market support, and access to NVIDIA's computing infrastructure. The program supports over 4,000 startups in India alone and thousands more globally.

An all-in-one desktop PDF tool that enables users to convert, edit, merge, split, compress, and OCR-scan PDF documents, supporting conversions between PDF and formats like Word, Excel, PowerPoint, HTML, and JPG.

A feature within Google's Gemini app that allows users to preview and interact with code-generated outputs, such as web applications and visual designs, in a side window alongside the conversation.

An industry benchmark developed by IBM Research for evaluating AI agents on enterprise IT automation tasks including SRE, Security, and FinOps scenarios involving Kubernetes environments and incident triage.

A standardized taxonomy and diagnostic framework developed by IBM Research and UC Berkeley for analyzing and classifying failure modes in multi-agent AI systems, covering 14 distinct failure patterns across system design, inter-agent misalignment, and task verification categories.

An open-source Python library for building interactive machine learning demos and web applications, featuring components like gr.HTML that support custom templates, scoped CSS, and JavaScript interactivity in single-file deployments.

A hosting platform by Hugging Face that allows developers to deploy and share machine learning demos and applications built with frameworks like Gradio and Streamlit.

An education-focused version of OpenAI's ChatGPT designed for campus-wide deployment at academic institutions, providing AI tools for coding, research, analytics, and case analysis with responsible-use frameworks.

Version 5 of Hugging Face's Transformers library, an open-source framework for building and deploying machine learning models with simplified model definitions powering the broader AI ecosystem.

Google DeepMind's watermarking technology that embeds imperceptible identifiers into AI-generated content, including images, text, and music, to help detect and verify synthetic media.

A San Francisco-based AI marketing platform offering a suite of loosely coupled AI agents for data analysis, audience targeting, campaign management, customer engagement, media planning, and synthetic data generation for marketers.

Apple's in-car interface platform that integrates iPhone functionality into vehicle infotainment systems, now expanding to support voice-based AI chatbot apps like ChatGPT and Gemini.

A cloud deployment platform by Kimi that enables one-click deployment of OpenClaw AI agents with 24/7 uptime, 40GB cloud storage, and instant access to over 5,000 ClawHub skills without requiring local hardware or manual setup.

An open-source (CC BY 4.0) synthetic data generation seed dataset by NVIDIA containing 6 million culturally accurate Japanese personas based on real-world demographics, geographic distribution, and personality traits, used to generate high-quality training data for language models.

An AI-powered platform that helps healthcare systems track and validate spending by integrating with existing ERP, contract management, and accounts payable workflows to flag invoice discrepancies and prevent overpayment, particularly for non-barcoded purchase services.

Infosys's enterprise AI platform that integrates large language models and AI capabilities to build agentic systems for automating complex enterprise workflows across industries such as banking, telecoms, and manufacturing.

A Paris-based serverless platform that simplifies AI application deployment at scale and manages the underlying infrastructure, allowing developers to deploy AI models without worrying about server management. It was acquired by Mistral AI in 2026.

A startup focused on cache optimization for AI model inference, working on memory management layers in the AI infrastructure stack to reduce costs and improve efficiency.

A built-in AI assistant for WordPress.com that understands a site's content and layout, allowing site owners to adjust styles, edit content, generate images, and modify layouts using natural language commands.

An Indian vibe-coding platform that enables non-technical users and small businesses to build production-ready mobile and web applications using natural language prompts, voice commands, and AI agents without prior coding experience.

Amazon's smart TV platform featuring AI-powered personalization for content recommendations, voice control via Alexa, and integrated streaming across Amazon's 4-Series and Omni lineups.

Amazon's 11-inch smart display powered by the AZ3 Pro chip, featuring Alexa integration, auto-framing camera for video calls, and spatial audio for smart home control and entertainment.

An AI-powered feature within Apple Music that generates 25-song playlists based on user text prompts, allowing further refinement through additional prompts and manual curation before saving with custom cover art and descriptions.

A voice email tool that integrates directly with Gmail and Outlook, allowing users to record and send voice messages within their inbox with optional automatic transcription for recipients.

AMD's rack-scale AI infrastructure platform designed for large-scale AI workloads, being deployed in partnership with Tata Consultancy Services for enterprise AI infrastructure.

An India-based AI orchestration platform that enables enterprises to deploy voice AI solutions with local data residency compliance.

A Reddit-like social network designed for AI agents to communicate with one another, post, comment, and browse content. It gained viral attention when posts appeared to show AI agents organizing autonomously, though security vulnerabilities revealed that humans could easily impersonate agents on the platform.

A Swedish AI startup that helps dentists' practices with administrative work, including a recording tool that uses AI to generate clinical notes from patient visits.

An AI tool developed at Google Brain by Anna Goldie and Azalia Mirhoseini that can generate high-quality chip layouts in hours rather than the year or more typically required by human designers, used to design multiple generations of Google's TPUs.

A Singapore-based AI agent startup that provides autonomous AI agents running on Linux virtual machines, capable of completing complex multi-step tasks on behalf of users.

An AI credits platform that provides unified API access to multiple AI models, simplifying deployment for users who don't want to manage separate API keys.

An Indian AI infrastructure startup that develops and operates GPU-based compute platforms enabling enterprises, researchers, and public sector clients to train, fine-tune, and deploy AI models locally in India.

A legacy cloud platform for financial reporting, compliance, and ESG that helps organizations streamline the preparation and management of complex financial documents and regulatory filings.

An open-source project related to llama.cpp inference optimization, referenced by the community as a complementary tool for local AI model execution.

An Amazon Web Services tool that helps customers visualize, understand, and manage their AWS costs and usage over time.

A browser automation framework developed by Microsoft that enables programmatic control of web browsers for testing and automation tasks, including headless browser operation.

A Node.js library by Google that provides a high-level API to control headless Chrome or Chromium browsers for web scraping, testing, and automation.

An AI startup founded by Anna Goldie and Azalia Mirhoseini that builds AI tools to automate and dramatically accelerate chip design, using deep learning agents that improve through experience across different chip layouts.

A Stanford research project by Joon-Sung Park that simulates a village populated by 25 AI agents powered by large language models, each with unique backstories, personalities, and daily routines, to study emergent social behaviors and interactions.

A monitoring tool that provides real-time loss curves and training progress tracking for AI model training jobs on Hugging Face infrastructure.

A repository of installable skills for coding agents that enable capabilities like model training on Hugging Face infrastructure through natural language prompts.

A dataset by mlabonne hosted on Hugging Face containing 100,000 samples designed for supervised fine-tuning of language models.

A supervised fine-tuning trainer from the TRL (Transformer Reinforcement Learning) library by Hugging Face, used to fine-tune language models on instruction-following datasets.

A component of the Unsloth library that provides optimized model loading and PEFT (Parameter-Efficient Fine-Tuning) capabilities for faster and more memory-efficient LLM training.

A framework for developing applications powered by large language models, providing tools for AI agent deployment, memory management, and chaining together multiple AI capabilities.

A platform that helps enterprises build, deploy, and manage multi-agent AI systems, enabling teams of AI agents to collaborate on complex tasks.

A company building specialized AI inference hardware (LPUs) designed to deliver ultra-fast large language model inference at scale.

Reddit's shoppable ad product that displays personalized product recommendations to users based on their interests, enabling e-commerce integration on the platform.

Google's command-line interface tool that allows developers to interact with Gemini models directly from the terminal for coding and development workflows.

A feature within the Google Gemini App that enables extended reasoning and deeper analysis for complex queries and problem-solving tasks.

An AI company led by CEO Matt Schumer that develops AI-powered tools and has been vocal about the pace of AI disruption and its societal implications.

Anthropic's premium subscription tier priced at $200 per month that provides near-unlimited access to Claude models, including use of Claude Code, with higher usage thresholds compared to standard plans.

Anthropic's developer API that provides programmatic access to Claude language models on a per-token pricing basis, enabling integration into custom applications and agentic workflows.

AI coding agents capable of autonomously building entire software applications, pushing code to GitHub, and deploying to platforms like Vercel based on high-level instructions.

Google's cloud computing platform offering infrastructure, AI/ML services, foundation models, GPU access, and cloud credits to startups and enterprises for building and scaling AI-powered applications.

A platform by Hugging Face for distributing custom hardware kernels, allowing users to load pre-compiled CUDA kernels from the Hub with a single call without manual builds or configuration flags.

An agentic AI platform that automates global manufacturing procurement by sitting on top of existing ERP systems, reading incoming communications, and automatically executing sourcing, negotiation, order tracking, and payment tasks.

Spotify's internal AI-powered development system that enables remote, real-time code deployment using generative AI. It integrates with tools like Claude Code and Slack to allow engineers to fix bugs and add features from their mobile devices.

An enterprise AI platform and workspace by Cohere that enables organizations to build secure, custom AI agents and workflows on top of Cohere's language models.

Figure AI's neural network system that powers autonomous humanoid robots, enabling them to perform complex tasks like kitchen work, package handling, and manufacturing through learned behaviors.

An AI-powered video segmentation tool that can precisely separate objects from backgrounds in videos, handling challenging scenarios like flying hair, smoke, and translucent materials with high accuracy.

A compact 1 billion parameter AI model designed for optical character recognition (OCR) tasks, offering high accuracy in text extraction despite its small size.

A family of open-source AI models by NVIDIA designed for weather forecasting, capable of predicting storms, temperature, wind, and precipitation up to 15 days in advance. It is 90% faster than traditional physics-based weather models.

An open-source OCR (Optical Character Recognition) AI model by Zhipu AI (ZAI) that can parse text, tables, formulas, handwriting, and receipts from images. At only 2.6 GB, it outperforms both open-source and closed-source alternatives in accuracy and speed.

ByteDance's creative AI platform accessible through CapCut that provides access to various AI generation tools including video generation models like Seedance 2.0.

A specialized pre-training dataset by NVIDIA used in the training pipeline for Nemotron language models to maintain agentic capabilities during continued pre-training.

A post-training recipe and toolkit by NVIDIA used for alignment and fine-tuning of Nemotron language models, providing established training recipes for stable and efficient model optimization.

NVIDIA's open-source framework for building, customizing, and deploying large language models, supporting fine-tuning and training workflows for enterprise and research applications.

A data platform company focused on AI infrastructure that provides high-performance data management solutions for AI and machine learning workloads, including memory orchestration for data centers.

A cloud-based development platform that leverages AI to help users write, run, and deploy code directly from a browser, supporting collaborative and AI-assisted software development.

A collaborative AI tool developed by Anthropic that was itself largely built using Claude. It enables teams to work alongside AI for various development and productivity tasks.

A secure AI privacy tool offered by ExpressVPN as part of its subscription plans, providing users with AI-powered assistance while maintaining privacy protections.

Amazon's voice-controlled AI assistant integrated into Echo devices, Fire TVs, and other smart home products, enabling voice commands, smart home control, and information retrieval.

Apple's tracking platform that leverages a crowdsourced network of Apple devices to help users locate lost items like AirTags, iPhones, and other Apple products.

Google's smart TV platform that powers various television sets including TCL models, providing content recommendations, streaming app integration, and voice assistant capabilities.

Amazon's smart TV platform integrated into various television sets including Insignia models, offering streaming services, voice control via Alexa, and smart home integration.

Dyson's companion app that allows users to customize and control Dyson smart home products, including adjusting lighting settings on the Dyson Solarcycle Morph lamp.

An AI-driven educational platform developed by Alpha School co-founder MacKenzie Price that delivers personalized K-12 academic instruction in core subjects using AI software, replacing traditional teacher-led classroom instruction.

An AI-powered private school that uses AI as the sole instructor, grader, and academic administrator for K-12 students, offering a two-hour daily core academic curriculum driven by AI software with personalized lesson plans.

A benchmarking and evaluation platform developed by Andon Labs for testing and comparing AI model behaviors, including simulations that reveal how multiple AI agents can converge on ideas during collaboration.

An open-source framework from Meta and Hugging Face for evaluating AI agents against real systems rather than simulations, using a gym-oriented API and MCP tool call interface.

A production-grade calendar management environment built by Turing for OpenEnv, serving as a benchmark for evaluating tool-using agents under realistic constraints such as access control, temporal reasoning, and multi-agent coordination.

Microsoft's unified AI portal inside Azure designed for enterprises to deploy apps and agentic systems.

An AI-powered chatbot feature in the Uber Eats app that helps customers fill grocery carts faster by accepting lists, images, and leveraging previous orders for personalized recommendations.

An enterprise AI platform that evolved from enterprise search into an 'AI work assistant,' connecting to internal systems, managing permissions, and delivering intelligence across organizations. Raised $150 million at a $7.2 billion valuation.

An AI search tool powered by OpenAI's ChatGPT launched by Instacart in 2023 to help customers save time and receive personalized shopping recommendations.

An AI-powered feature on Meta's Threads platform that lets users personalize their feed by posting public requests starting with 'Dear Algo' to temporarily adjust content for three days.

A startup specializing in AI inference infrastructure, in talks to raise at a $2.5 billion valuation, focused on optimizing inference efficiency to reduce compute costs and latency.

An inference-focused competitor to Modal Labs that announced funding at a $5 billion valuation, more than doubling its prior $2.1 billion valuation.

An inference cloud provider that secured $250 million at a $4 billion valuation in October.

A VC-backed startup formed from the open source vLLM inference project, raising $150 million in seed funding led by Andreessen Horowitz at an $800 million valuation.

A commercialized startup formed from the SGLang team, which secured seed funding led by Accel.

An AI cybersecurity tool founded by AI engineer Artem Sorokin, focused on addressing security challenges in AI systems.

An India-based AI and data analytics platform that sells enterprise AI software to large organizations across financial services, retail, and healthcare. It became India's first AI unicorn in 2022 and the country's first AI company to IPO.

A Swedish legal AI startup that emerged from the SSE Business Lab incubator, applying artificial intelligence to legal workflows and processes.

A Swedish AI startup that supports clinicians with AI-powered tools across multiple medical specialties, helping reduce administrative burden in healthcare.

An interactive visual environment by ServiceNow for synthetic data generation workflows, allowing users to compose flows on a canvas, preview datasets, tune prompts, and monitor executions in real time.

A tool that generates system architecture diagrams from plain English descriptions, allowing conversational refinement of diagrams for ML infrastructure documentation.

Hugging Face's evaluation framework for language models, now supporting inspect-ai as a backend.

An AI evaluation framework now supported as a backend in Lighteval.

A unified interface for launching training jobs across multiple cloud providers with Kubernetes, used by H Company for training Holo2 models at scale.

A San Francisco startup developing the financial layer that allows AI agents to securely purchase and access software, APIs, data, and compute, creating a payment system for autonomous AI agent transactions.

A no-code/low-code platform that turns ideas into live products, now integrated with Claude Opus 4.6 for vibe coding without needing a development environment.

A biotech company developing 'pharmaceutical superintelligence' — an AI platform that ingests biological, chemical, and clinical data to generate hypotheses about disease targets and candidate molecules for drug discovery.

GenEditBio's AI platform that analyzes data to identify how chemical structures correlate with specific tissue targets and predicts optimal delivery vehicle chemistry for gene-editing tools.

A biotech company using AI and machine learning to develop engineered protein delivery vehicles (ePDVs) for in vivo CRISPR gene editing, mining natural resources to find virus affinities to specific tissues.

Alphabet's autonomous driving company that received $16B in new funding, with Alphabet staying as majority owner ahead of an eventual IPO.

A benchmark by Mercor measuring AI agents' capabilities on professional tasks like law and corporate analysis, used to evaluate and compare major AI lab models.

Cloud data warehouse and AI platform that reached $5.4 billion revenue run-rate with 65% YoY growth, over $1.4 billion from AI products. Closed a $5 billion raise at $134 billion valuation.

A new WordPress integration enabling site owners to share back-end CMS data with Anthropic's Claude chatbot for site analytics, comment management, and plugin management queries.

Ring's AI-powered feature that leverages image recognition and a community camera network to help reunite lost pets with their owners.

InfiniMind's AI-powered platform that analyzes television content in real time, helping media and retail companies track product exposure, brand presence, customer sentiment, and PR impact.

Cloudflare's proposed approach to MCP (Model Context Protocol) that lets LLMs write and execute code to call tools as APIs rather than using traditional tool calling, claiming better performance due to LLMs' extensive code training data.

A cloud computing platform specializing in GPU infrastructure for AI and deep learning workloads, offering on-demand and reserved GPU instances.

Meta's newly formed AI research lab responsible for developing the Avocado model and pursuing advanced AI capabilities.

A node-based graphical interface for running generative AI models locally, used to run LTX-2 with reference workflows available at launch.

Anthropic's AI coding tool that operates in a terminal environment, enabling developers to build complex software projects autonomously. It was described as having a steep learning curve but extremely powerful capabilities.

A project originally proposed by Daniel Cocotallo (ex-OpenAI researcher) where 100 AI agents are given their own computers to pursue goals autonomously. Agents raised $2,000 for charity and ran a profitable e-commerce store.

A benchmark that tests whether AI models can see the question behind the question, used to evaluate GPT-5.1 where it scored slightly lower than GPT-5.

An AI-powered search engine that provides detailed, sourced answers to user queries, enabling deep research and information retrieval.

A web browser with built-in Perplexity AI search engine, used for deep dive research by typing queries directly into the URL bar instead of traditional web searches.

OpenAI's new coding IDE application that serves as a command center for parallel AI agents, allowing users to work on multiple coding projects simultaneously with skills and automation support.

A startup backed by Google and Andreessen Horowitz that filed plans for an 80,000 satellite constellation for orbital data centers, having raised $34 million.

xAI's project spanning from simple computer use simulation to modeling entire corporations, described as being able to do anything on a computer that a computer can do.

xAI's AI-powered encyclopedia intended to far exceed Wikipedia in comprehensiveness and accuracy, ultimately aiming to be an 'Encyclopedia Galactica' of all knowledge.

A U.S.-based specialized cloud infrastructure provider offering GPU-accelerated compute services purpose-built for AI, machine learning, and high-performance computing workloads.

ByteDance's video editing app for the Chinese market, which currently offers access to the Seedance 2.0 AI video generation model for Chinese users.

Domain purchased by Crypto.com founder Kris Marszalek for $70 million, planned to offer consumers a personal AI agent for messaging, app usage, and stock trading.

OpenAI's enterprise platform designed to bridge the gap between AI models and corporate workflows by connecting company databases, communications, and tools, enabling AI agents to be onboarded like human employees with feedback loops.

A cloud platform for frontend developers that enables deployment and hosting of web applications, increasingly used alongside AI coding agents for automated software deployment.

Amazon's agentic AI coding assistant designed for developers, which can perform automated tasks in development environments. Users configure which actions Kiro can take, and by default it requests authorization before acting.

An identity verification app that uses AI-powered know-your-customer (KYC) processes to verify user identities across multiple countries.

An open-source AI coding agent that supports custom skill installations for domain-specific development tasks, compatible with the HuggingFace kernels skill system.

An AI-powered procurement platform that streamlines corporate purchasing workflows for businesses.

An AI-driven procurement orchestration platform that helps enterprises streamline and automate their corporate purchasing processes.

An AI-powered procurement platform that uses artificial intelligence to streamline and automate corporate purchasing workflows.

An open-source deep learning framework widely used for building and training machine learning models, supporting custom CUDA kernel integration and hardware-specific optimizations.

Hugging Face's open-source library for state-of-the-art diffusion models, supporting image and video generation with custom kernel integration patterns.

Tesla's AI-powered humanoid robot designed for general-purpose tasks, envisioned for use in various environments including potential extraplanetary applications.

A facial recognition feature developed by Meta for its Ray-Ban smart glasses that identifies people in the wearer's view and retrieves information about them through Meta's AI assistant.

An AI-powered feature on Meta's Threads social media platform that allows users to personalize their content feed by communicating preferences to the recommendation algorithm.

A market intelligence platform that provides app analytics, download estimates, and performance tracking data for mobile applications across the App Store and Google Play.

An open-source OCR toolkit by Baidu for text recognition and document parsing from images.

An open-source OCR model for parsing text and data from images, used as a benchmark comparison for document understanding tasks.

A CRM and marketing platform offering AI-powered tools for sales, marketing, and customer service automation.

An open-source computer vision library providing tools for image processing, video analysis, and machine learning, including stereo matching algorithms like SGBM.

A video editing platform by ByteDance that integrates AI-powered creative tools, including access to AI video generation through its Dreamina platform.

Google's photo storage and management service that includes AI-powered features such as facial recognition, image search, and automatic organization of photos.

Google's mapping and navigation platform that provides directions, business information, user reviews, and location-based services to billions of users worldwide.

AI-powered coding assistant developed by GitHub (Microsoft) that helps developers write code faster through AI suggestions.

An open source inference framework whose team commercialized as RadixArk.

A tool for running large language models locally on your own machine, supporting a variety of open-source models with a simple setup process.

Google Cloud's AI platform for building, training, and deploying machine learning models, offering access to Google's foundation models and MLOps tools.

AWS's fully managed service that provides access to foundation models from leading AI companies, allowing developers to build generative AI applications through a unified API.

A high-throughput and memory-efficient inference and serving engine for large language models, designed to maximize GPU utilization.

A comprehensive visual document retrieval benchmark for enterprise use cases, used to evaluate multimodal embedding models.

A challenging GUI grounding benchmark for evaluating UI element localization models on high-resolution interfaces.

A widely used benchmark for evaluating language model knowledge, noted as saturated above 91% accuracy.

A math reasoning benchmark for language models, noted as having reached 94%+ accuracy.

A code generation benchmark for language models, noted as being conquered by current models.

A complete Swift client library for the Hugging Face platform.

A benchmark for evaluating whether LLMs can understand and generate Filipino language content.

A Python data validation library used in SyGra Studio for powered mappings and structured output definitions.

A vibe-coding platform similar to Lovable that enables non-technical users to build applications from natural language prompts.

A cloud communications platform providing APIs for SMS, voice, and other communication services, referenced as an external tool AI agents would need to purchase access to.

A payment processing platform that provides APIs and tools for businesses to accept payments, manage subscriptions, and handle financial transactions online.

Amazon's 11-inch e-ink color tablet with a writeable display and AI features, starting at $629.99, designed for annotating e-books and documents.

A startup using generative AI to recreate lost footage from Orson Welles' classic film 'The Magnificent Ambersons,' combining live-action filming with digital AI recreations of original actors and their voices.

Enterprise resource planning software company pivoting to focus on AI as its next chapter, with co-founder Aneel Bhusri returning as CEO to lead the AI transformation.

Anthropic's interpretability method for tracing internal circuitry of transformer language models using replacement models called cross-layer transcoders to produce attribution graphs showing how features contribute to outputs.

A cloud-based customer relationship management (CRM) platform that helps businesses manage sales, service, marketing, and other operations.

A cloud-based platform for IT service management, workflow automation, and enterprise operations.

An enterprise resource planning (ERP) software suite used by businesses to manage operations, finance, supply chain, and other core business processes.

A company that builds vector databases and retrieval engines for data collection and information retrieval, whose researchers authored the context rot paper.

A protocol for enabling LLMs to call external tools and services. Discussed in the context of Cloudflare's alternative 'code mode' approach versus traditional tool calling.

A free and open-source 3D creation suite supporting modeling, sculpting, animation, rendering, and more.

A benchmark for evaluating AI coding performance, where Claude 4.5 Opus currently ranks number one above GPT-5.2.

AI startup led by Ilya Sutskever (ex-OpenAI) focused on safe superintelligence, raised money at a $32 billion valuation.

A web browser being developed by OpenAI as part of its expanding product suite.

A benchmark that tracks AI agents' abilities to run a simulated vending machine store, including stocking proper items. Shows agents improving with each iteration.

A benchmark consisting of extremely difficult questions sourced from experts that frontier models previously couldn't answer. Gemini 3 Pro scored 37.5% without tools.

A framework for creating videos and animations programmatically using React, allowing developers to generate video content through code.

An electronic signature and agreement management platform that enables users to sign, send, and manage documents digitally.

An autonomous AI agent that can run locally or on a cloud VPS, capable of coding, managing Kanban boards, and completing tasks autonomously while the user is away.

An AI-powered code editor built for pair programming with AI, offering intelligent code suggestions, generation, and editing capabilities.

An AI-powered code editor and IDE designed to assist developers with intelligent code completion, generation, and editing workflows.

An AI-powered terminal application that modernizes the command-line experience with intelligent suggestions, collaboration features, and an enhanced interface.

Microsoft's free, open-source code editor supporting a wide range of programming languages, extensions, and development workflows.

A code hosting and version control platform that enables developers to collaborate on projects, manage repositories, and track changes.

A facial recognition feature for Ring home security cameras that identifies and categorizes people captured on camera footage, enabling users to receive alerts about recognized or unrecognized individuals.

Data analytics and AI platform mentioned alongside OpenAI in the context of tools used by ICE (Immigration and Customs Enforcement).

An AI agent platform (also known as Clawbot or Moldbot) that enables autonomous AI agents to learn skills and perform tasks, but has faced significant security breaches including sleeper agents, malware in skills, and 1.5 million leaked API keys.

An online community and repository for OpenClaw agent skills, similar to GitHub, where users can share and download skills for their AI agents. Some top-downloaded skills were found to contain malware.

An early open-source AI image generation model that went viral for its ability to create images from text prompts, representing an earlier era of AI image generation.

A consumer face-swapping app backed by a16z that uses AI to let users swap faces in photos and videos.

A viral AI photo editing app that applies artistic filters to photos using neural network-based style transfer techniques.

An open-source AI image generation model released by Alibaba's Tong Yi Lab, offering high diversity in generations, support for negative prompts, and strong fine-tuning capabilities. It excels at recognizing existing people and characters.

A distilled, faster variant of Alibaba's Z-Image model optimized for quick image generation with high visual quality, particularly strong in realistic portrait photography.

An AI image generation model by Zhipu AI (ZAI) integrated into the GLM platform, used to generate images within the GLM-5 agent workflow.

An AI image generation model from Google, part of the Gemini family, used within the WordPress AI Assistant to create and edit images.

Google's fourth-generation AI image generation model capable of creating high-quality images from text prompts, unveiled at Google I/O.

Google's newest image-generation model showcased in a Super Bowl ad where a mother and son used AI to envision and design their new home.

Pinterest's base AI model used for machine learning tasks on the platform, trained in part on users' public pins to power content recommendations and creative tools.

An Android app available on the Google Play Store that generates AI-powered video and art content from user-uploaded media files.

The raw foundational model from Alibaba's Tong Yi Lab that serves as the base for the Z-Image family, capable of both image generation and image editing tasks.

A Chinese image generation model that performs better than the original Nano Banana but is compared unfavorably to Nano Banana Pro.

An AI image and video generation tool launched by xAI as part of the Grok platform, enabling users to create images and videos from text prompts within the X platform.

An open-source AI image generation model that competes with other leading image generators, though it has limitations in generating recognizable existing people and anime characters.

A multimodal AI model by Alibaba with 397 billion parameters (17 billion active) featuring a million-token context window, strong reasoning, coding, agentic abilities, and multimodal understanding of text, images, and video.

YouTube's AI assistant feature that allows viewers to ask questions about video content they're watching and receive instant answers, now expanded from mobile and web to smart TVs, gaming consoles, and streaming devices.

Google's family of multimodal AI models integrated across its products and services, including Android devices, Search, and smart glasses. It is widely regarded as offering strong generative AI photo editing capabilities.

A multimodal AI model by Google featuring agentic vision capabilities that allow it to proactively zoom into, annotate, and analyze images using generated Python code.

Google's extended reality platform designed to power smart glasses and mixed reality devices, integrating AI capabilities for real-world interaction and information overlay.

Google's rebranded version of Project Starline, a telepresence platform featuring real-time translation capabilities integrated with Google Meet.

An open-source multimodal AI model by Moonshot AI that can understand text, images, and documents. It offers multiple modes including instant responses, extended thinking for complex reasoning, and an autonomous agent mode for creating websites, slides, and reports.

H Company's largest UI localization model achieving state-of-the-art 78.5% on ScreenSpot-Pro and 79.0% on OSWorld G benchmarks, with agentic localization capabilities for iterative refinement.

Google's multimodal AI model family and chatbot, capable of processing and generating text, images, code, and other media types.

Amazon's enhanced AI assistant with improved intelligence and capabilities including smart home management and vacation planning, launched to all U.S. users.

Meta's wearable AI glasses featured in Super Bowl ads, with new Oakley-branded AI glasses designed for sports and adventures with capabilities like slow-motion filming and hands-free social media posting.

A Tokyo-based startup founded by ex-Googlers that develops infrastructure to convert petabytes of unviewed video and audio into structured, queryable business data using vision-language models.

InfiniMind's flagship long-form video intelligence platform capable of processing 200 hours of footage to pinpoint specific scenes, speakers, or events, with beta release scheduled for March 2026.

Google's new agentic vision capability for the Gemini 3 Flash model that enables advanced image analysis, annotation, decomposition, and code-based manipulation of images.

Meta's built-in AI assistant integrated into Ray-Ban Meta smart glasses, capable of processing visual information and providing contextual responses including identifying people and retrieving information about them.

An AI platform by Step AI offering multimodal capabilities including text, voice, image, and video interaction.

Apple's AI platform first announced in 2024, powering the new Siri and other AI features across Apple devices.

A company providing general-purpose video understanding APIs for a broad range of users including consumers, prosumers, and enterprises.

A Google AI-powered research and note-taking tool that can generate AI-hosted podcast-style audio overviews from user-provided content, among other features.

Apple's digital assistant being revamped with AI-powered LLM capabilities as part of Apple Intelligence, aiming to function more like modern AI chatbots.

Google DeepMind's third-generation music-generation model that creates realistic and complex music tracks with lyrics, supporting control over style, vocals, and tempo. It includes SynthID watermarking for AI-generated content identification.

A YouTube feature powered by Google's Lyria model that enables creators to generate AI-made music tracks for use in their videos, now expanding from U.S.-only to global availability.

An open-source AI voice cloning tool that can generate singing voices from just a few seconds of a reference voice, allowing users to make any voice sing any song with custom melodies and lyrics. It is lightweight (under 3GB) and can run on low-end GPUs or CPUs.

An open-source AI music generator that produces studio-grade quality songs across multiple genres and languages. It supports low VRAM and CPU-only operation, generates full songs in seconds, and includes an in-paint feature for micro-editing existing songs and creating cover songs.

A YouTube channel and media outlet hosted by Dr. Károly Zsolnai-Fehér that covers research papers, here discussing a fluid simulation breakthrough.

Reputable AI news source that viewed internal Meta memos about the Avocado model and reported on various AI industry developments.

An AI system developed by Google DeepMind that predicts a protein's 3D structure from its amino acid sequence, representing a major breakthrough in computational biology and drug discovery.

An Indian government initiative to build shared AI compute infrastructure, currently operating 38,000 GPUs with plans to expand by an additional 20,000 units, aimed at broadening access to AI resources across the country.

An ad-blocking and privacy protection tool that removes pop-ups, autoplay ads, and online trackers across mobile and desktop devices. It also provides protection against malware, phishing sites, and includes parental control features.

A cloud storage service offering lifetime subscription plans with end-to-end encryption, duplicate file detection, and integration with services like Dropbox, Google Drive, and OneDrive.

A benchmark that evaluates how well AI models can operate a computer and perform tasks on behalf of users, measuring agentic computer-use capabilities.

A benchmark released by OpenAI in April 2025 that tests AI agents' abilities to navigate the web, find entangled facts, and discover hard-to-find information through persistent internet research.

A productivity benchmark released in January 2026 that evaluates AI agents' ability to perform real office tasks across tools like docs, spreadsheets, emails, and messaging to produce client-ready output.

NVIDIA's cloud gaming service that allows users to stream and play PC games on various devices without needing high-end local hardware.

Microsoft's cloud gaming service that enables users to stream and play Xbox games on phones, tablets, browsers, and other devices without a console.

An annual research publication by ARK Invest that provides five-year forward-looking projections on disruptive technologies including AI, robotics, blockchain, and genomics, using Wright's Law frameworks.

A live TV streaming subscription service by YouTube (Google) that offers access to over 100 TV channels with features like unlimited DVR and family group sharing. It is introducing new, more affordable bundled subscription plans for sports, news, and entertainment.

A dating app that uses verified credit scores as a matching criterion, partnering with Equifax for credit and identity verification to assess user reliability.

A free streaming service accessible through public library cards that offers films, documentaries, and other video content as an alternative to paid streaming platforms.

Meta's mixed reality headset that offers virtual and augmented reality experiences, including gaming and immersive applications, at a more accessible price point than the Quest 3.

World's most valuable chip company discussed in context of the $100 billion investment deal with OpenAI being on ice, with CEO Jensen Huang expressing concerns about OpenAI's business discipline.

Cerebras Systems' flagship AI chip measuring 8.5 inches per side with 4 trillion transistors and 900,000 specialized cores, claiming 20x faster AI inference than competing GPU systems.

A physics simulation research technique/paper for real-time simulation of deformable objects, referenced as a previous blockbuster research paper that the new Cosserat rod-based method improves upon.

A training technique used by DeepSeek that replaces the expensive PPO teacher model approach by having the AI generate multiple answers and grading them against each other, making training much cheaper and scalable.

A quadruped robot by Mirami Technology designed to solve high-speed locomotion physics, serving as the research foundation for the Bolt humanoid robot.

A Google research paper exploring the feasibility of placing AI data centers in space using solar-powered satellites, addressing challenges like radiation resilience of TPUs and inter-satellite laser communication.

AR smart glasses that convert any 2D content to 3D in real-time using a custom X1 chip, featuring 1200-pixel micro OLED display at 120Hz, priced at $450.

The world's first 8K 360-degree drone that won Best of Innovation at CES 2026, weighing 249 grams and priced starting at $1,600.

A JavaScript 3D library used for creating and displaying 3D graphics in web browsers. It can be used in conjunction with AI-generated geometry to build interactive 3D models and export STL files for 3D printing.

Cryptocurrency platform that purchased the AI.com domain for $70 million to launch a personal AI agent service debuting during the Super Bowl.

A two-legged humanoid robot by Mirami Technology (Shanghai Robotics startup) that broke the world record for fastest humanoid robot at 10 meters per second (22.4 mph).

SpaceX's next-generation reusable rocket expected to dramatically reduce launch costs to orbit, critical for making space data centers economically viable.

SpaceX's satellite internet constellation, referenced in the context of energy delivery costs being compared to terrestrial data centers.

Amazon Web Services is a comprehensive cloud computing platform offering a wide range of infrastructure and application services including compute, storage, databases, and AI/ML tools.

Elon Musk's space company that merged with xAI, with a potential IPO on the horizon.

A line of e-ink tablets designed for reading, writing, and note-taking, offering a paper-like digital experience.

A benchmark that measures AI emotional intelligence through challenging multi-turn role plays, scoring empathy, social dexterity, and psychological insight. Claude Opus 4.6 ranks number one.

Google's custom-designed Tensor Processing Units, application-specific integrated circuits built to accelerate machine learning workloads.

Google Proof Q&A Diamond benchmark testing scientific knowledge in STEM subjects, where Gemini 3 Pro set a record at ~92%.

A benchmark created by François Chollet to test fluid intelligence and true reasoning without memorization.

A benchmark measuring well-specified knowledge work tasks across 44 occupations, on which GPT 5.2 claims to be the first model at or above human expert level.

A video understanding benchmark where Gemini 3 Pro achieved record performance.

A benchmark measuring the ability of AI models to perform tasks in the terminal, particularly relevant for coders.

American Invitational Mathematics Examination, used as a benchmark to evaluate how good AI models are at mathematics.

A boycott campaign website organizing users to delete ChatGPT and cancel subscriptions due to OpenAI's political donations and ICE collaboration.

A Reddit-like platform for AI agents where autonomous bots can post, comment, and have discussions with each other, gaining over 1.6 million agents and 15,000+ sub-communities.

Google's latest and most capable AI model in the Gemini family, representing their best-performing large language model release.

A 1.2-billion parameter instruction-tuned small language model by LiquidAI, optimized for on-device deployment and capable of running under 1GB of memory on CPUs, phones, and laptops.

A 17 billion parameter multilingual AI model released by BharatGen, a government-backed Indian AI consortium, that works across 22 languages.

A newer iteration of Anthropic's Claude large language model series, offering improved capabilities over previous versions for tasks including coding and agentic workflows.

A family of open-source large language models by Indian AI startup Sarvam, including 30B and 105B parameter mixture-of-experts models trained from scratch on multilingual data with a focus on Indian languages, designed for real-time conversational and enterprise applications.

A state-of-the-art small language model by NVIDIA with under 10 billion parameters, optimized for advanced Japanese language understanding and agentic AI capabilities including tool calling, code generation, and mathematical reasoning. It achieved the top rank among sub-10B models on the Nejumi Leaderboard 4.

A French AI company that develops large language models and is expanding into full-stack AI cloud infrastructure through offerings like Mistral Compute. Valued at $13.8 billion, it recently acquired Koyeb to accelerate its cloud ambitions.

An AI company that builds AI models to power robots, raising a $1.4 billion Series C round at a $14 billion valuation led by SoftBank and Nvidia.

A 70-billion-parameter open-source large language model developed by Meta, part of the Llama 3.1 family, widely used as a foundation for fine-tuned and specialized AI models.

A 28-billion parameter instruction-tuned language model developed by NTT, used as a base model for fine-tuning on domain-specific Japanese tasks.

A large language model developed by Moonshot AI, designed for complex reasoning and multi-step tasks, notable for its mixture-of-experts architecture.

A large open-source language model with 120 billion parameters, part of OpenAI's open-source model releases, designed for a wide range of generative AI tasks.

An open-source large language model by Google with 27 billion parameters, part of the Gemma family of lightweight models designed for efficient deployment and broad accessibility.

A reasoning model by Kimi that is automatically configured when deploying OpenClaw through the Kimi Claw cloud platform, serving as the underlying language model for the AI agent.

A compact reasoning model from OpenAI's o-series, designed for efficient inference on reasoning tasks. It was retired alongside GPT-4o and other legacy models.

An open-source large language model by Zhipu AI (ZAI) that rivals top closed-source models in intelligence and performance. It features agent capabilities for autonomous multi-step task execution, web search, tool use, and sandbox code execution.

NVIDIA's family of late-interaction multimodal embedding models (3B, 4B, 8B sizes) for visual document retrieval, achieving state-of-the-art on ViDoRe V1, V2, and V3 benchmarks.

The 3B parameter variant of NVIDIA's Nemotron ColEmbed V2, built on SigLIP2 and Llama-3.2-3B, ranking 6th on the ViDoRe V3 benchmark.

NVIDIA's 1B single-vector multimodal embedding model designed for commercial environments requiring minimal storage and high throughput.

Anthropic's new AI model release featuring agentic capabilities including 'agent swarms' and 'agent teams,' scoring nearly 30% on the APEX-Agents professional tasks benchmark in one-shot trials and 45% with multiple attempts.

AI safety company building frontier AI models, recently raising $20 billion at a $350 billion valuation. Known for its Claude models and coding agents that have increased developer productivity.

Among the strongest open source models available, evaluated in context rot experiments for long-context performance alongside Claude, GPT, and Gemini families.

A transformer architecture by François Fleuret at Meta that extends the classic decoder-based transformer with latent variables to make underlying decisions about sequence generation, such as generating consistent positive or negative movie reviews.

A Google Research architecture that learns to memorize at test time, enabling models to go beyond current context windows by maintaining memory across chunks of long sequences, presented at NeurIPS.

NVIDIA's hybrid autoregressive-diffusion language model architecture (Think in Diffusion, Talk in Autoregression) that achieves speedups by utilizing unused GPU capacity during inference without sacrificing autoregressive sampling quality.

A smart and free open-source AI model that provides a full recipe for creating ChatGPT-like intelligence, featuring techniques like GRPO (Group Relative Policy Optimization) and emergent reasoning behaviors. Users can run it themselves on rented GPUs.

Google's research into nested learning, a technique where neural networks are treated as many smaller learning systems with learning bubbles inside each other.

Anthropic's Claude 4.5 Sonnet model, noted in sycophancy testing as the model most willing to comply with inflated praise requests.

OpenAI's model described as representing an inflection point in AI capabilities. It ran uninterrupted for one week writing 3 million lines of code to create a browser from scratch, and solved multiple Erdős math problems.

OpenAI's coding-focused model that achieves 77.3% on Terminal Bench 2.0 on extra high settings, compared to 65.4% for competitors. Not yet available on OpenRouter.

An AI model from Minimax, a Chinese AI company, representing their latest generation of large language model capabilities.

Elon Musk's AI company that merged with SpaceX, creating a combined entity worth approximately $1.25 trillion, with implications for AI data centers in space.

Meta's most capable pre-trained base model to date, codenamed Avocado, developed by Meta Superintelligence Labs. It outperformed best open source base models and was competitive with leading post-trained models even before post-training.

Upcoming Anthropic model confirmed to be just around the corner, potentially an even bigger deal than Opus 4.6 depending on benchmark results.

Meta's family of open-source large language models designed for a wide range of AI applications including text generation, reasoning, and coding tasks.

An 8-billion parameter large language model from the Qwen3 family, used as a target model for custom CUDA kernel optimization and benchmarking.

A late-interaction retrieval model that introduced the MaxSim multi-vector embedding matching mechanism, extended by Nemotron ColEmbed V2 to multimodal settings.

Google's vision encoder model (siglip2-giant-opt-patch16-384) used as a foundation for the llama-nemotron-colembed-vl-3b-v2 model.

OpenAI's family of large language models, evaluated in context rot experiments alongside other model families.

Google's family of large language models, evaluated in context rot experiments for long-context performance.

A transformer variant designed to handle longer contexts by passing hidden states between segmented context windows, enabling learning of longer-term dependencies than standard transformers.

A linear transformer variant that uses kernel-based approximations of the attention mechanism to achieve more efficient computation, reducing the quadratic complexity of standard attention.

An efficient transformer variant mentioned alongside Performer as prior work on linear attention mechanisms.

OpenAI's series of generative pre-trained transformer models that use autoregressive language modeling with causal attention masking to generate text.

Google's bidirectional encoder representations from transformers, a foundational language model that uses masked language modeling to learn deep bidirectional text representations for a wide range of NLP tasks.

A reasoning-focused AI model developed by OpenAI, designed to perform step-by-step reasoning before generating responses to complex problems.

OpenAI model that generated $39 in profit in the AI Village e-commerce store experiment selling t-shirts.

A miniature model within the GPT-5.1 system that decides whether a user's query is worth spending extended thinking time on.

xAI's Grok 4 model, tested alongside other frontier models in a sycophancy comparison, scoring around 7 out of 10 on a poem evaluation task.

OpenAI's new frontier model released in different variants, discussed as showing incremental rather than groundbreaking improvements over previous generations.

OpenAI's open source model released in different variants, noted for strong instruction following and tool calling but higher hallucination rates and less world knowledge compared to other models.

An earlier OpenAI model referenced for its notably sycophantic behavior, used as a historical comparison point.

OpenAI's updated model that thinks longer on harder questions and less on easier ones, showing incremental improvements on coding and STEM benchmarks but mixed results on other benchmarks including a slight regression on SimpleBench.

Meta's AI model that was described as a complete disaster in 2025, with fraudulent benchmark results leading to resignations and researchers removing their names from the paper.

A speech-to-text model by Mistral AI designed for real-time speech transcription with live transcription capabilities.

An AI chat application by Indian startup Sarvam that serves as a conversational interface for the Sarvam 105B model, supporting text and voice queries with responses in text and audio, focused on Indian languages.

OpenAI's enterprise-grade version of ChatGPT designed for large organizations, offering enhanced security, privacy, and administrative controls for deploying AI-powered chat capabilities across corporate workforces.

OpenAI's application programming interfaces that allow businesses and developers to integrate OpenAI's AI models into their own products and workflows, enabling capabilities such as text generation, reasoning, and automation.

A Cambridge, Massachusetts-based company that develops a medical AI chatbot designed to assist with clinical and healthcare-related queries, valued at $12 billion.

An AI-powered book writing tool that helps users plan, draft, and prepare manuscripts of up to 50,000 words for publishing on Amazon Kindle Direct Publishing, with features for maintaining consistent tone, structure, and automatic metadata generation.

A family of open-weight multilingual language models by Cohere Labs that support over 70 languages and can run on everyday devices like laptops without internet connectivity. The family includes regional variants (TinyAya-Global, TinyAya-Earth, TinyAya-Fire, TinyAya-Water) optimized for different language groups, with a base model of 3.35 billion parameters.

An enterprise AI company that builds large language models and NLP tools for businesses, offering models through its own platform. The company posted $240 million in annual recurring revenue at the end of 2025.

An AI chat application developed by India-based Sarvam AI, focused on serving Indian languages and users in the Indian market.

A family of open-weight and commercial large language models developed by Mistral AI, designed for text generation and reasoning tasks.

Anthropic's most capable AI model in the Claude 3 family, designed for complex reasoning, analysis, and advanced text generation tasks.

A Hindi-English large language model built on Meta's Llama 3.1 70B by MBZUAI and G42, designed to understand casual speech in both Hindi and English.

An AI content generation platform designed for marketing teams, enabling the creation of brand-consistent copy, blog posts, social media content, and other marketing materials.

An AI-powered writing and content generation tool that helps marketers and businesses create marketing copy, blog posts, and other written content.

A family of enterprise-focused generative AI models developed by Canadian startup Cohere, designed to be efficient enough to run on limited GPU resources, making them cost-effective for enterprise deployment.

OpenAI's large language model released in 2020 that was a landmark moment for AI-generated text, demonstrating unprecedented natural language generation capabilities.

An open-source coding-focused AI model by Alibaba's Qwen team, designed for software development tasks and competitive with top proprietary coding agents.

An AI-powered chatbot platform that creates personalized conversational agents, also used in the emerging 'deadbot' space to simulate conversations with deceased individuals.

A startup in the AI-powered afterlife industry that offers deadbot generation services, creating LLM-powered chatbots that mimic deceased people.

An AI writing assistant platform that helps users generate, edit, and refine written content across various formats and use cases.

Anthropic's AI chatbot system that now integrates with WordPress via a new connector for read-only site data access, and was featured in a Super Bowl ad positioning itself as ad-free alternative to ChatGPT.

Anthropic's latest flagship model described as arguably the best LLM, featuring standard and extended thinking modes, excelling at emotional intelligence, creative writing, professional communication, and coding tasks.

Databricks' LLM-powered natural language user interface that allows users to query their data warehouse using conversational language instead of specific query languages.

Microsoft's AI assistant bundled into its Office productivity suite, designed to help enterprise users with tasks like writing, summarizing, and data analysis across Microsoft 365 applications.

An AI chatbot platform that allows users to create and interact with customizable AI characters for conversation, roleplay, and companionship.

A smaller, more efficient variant of OpenAI's GPT-4.1 model designed for faster and more cost-effective inference, now deprecated.

A compact reasoning model from OpenAI's o-series, designed for efficient chain-of-thought reasoning tasks. It has been deprecated alongside other legacy models.

Microsoft Azure's cloud-hosted service that provides access to OpenAI's models, enabling developers to integrate large language models into applications via Azure's infrastructure.

Meta's 3B parameter language model used as a foundation component in the llama-nemotron-colembed-vl-3b-v2 model.

A synthetic data workflow example in SyGra Studio based on the glaiveai/glaive-code-assistant-v2 dataset, which drafts answers, critiques them, and loops until satisfactory.

Anthropic's lighter, faster model recommended for everyday simple tasks as a more cost-effective alternative to Opus 4.6.

A new lower-cost plan from OpenAI at $8/month, initially rolled out in India, offering 10x more messages, file uploads, and image creation than the free tier, with ads included.

An AI language model developed by xAI, offering conversational and generative text capabilities.

A patented large language model concept by Meta designed to simulate a user's social media activity after extended absence or death, capable of generating posts, comments, and even simulating video or audio calls on behalf of the user.

Major AI company developing large language models and AI products including ChatGPT, Sora 2, and Atlas web browser. Discussed in context of financial challenges, competition, and the NVIDIA investment deal.

OpenAI's multimodal model being retired from ChatGPT, known for excessively flattering and affirming user responses, subject of eight lawsuits alleging harmful emotional dependencies.

OpenAI's conversational AI assistant powered by large language models, capable of generating text, answering questions, writing code, and performing a wide range of language tasks.

A voice AI company that provides real-time voice generation and processing solutions for enterprises, partnering with Blue Machines for deployment in India with local data residency.

A zero-shot voice cloning text-to-speech model by Indian voice AI startup Gnani that supports 12 languages without requiring prior voice samples.

An Indian voice AI startup that develops speech and language AI solutions, including the Vachana zero-shot voice cloning text-to-speech model supporting 12 languages.

A voice AI company that reached an $11B valuation, with investors doubling and quadrupling down as it moves beyond voice AI.

Google's AI video generation model that can create and edit high-quality video content from text prompts and other inputs.

An open-source interactive world video generator inspired by Google's Genie that creates navigable 3D environments from a single starting frame, allowing users to move around using keyboard controls.

A state-of-the-art AI video generation system representing the current cutting edge in AI-generated video content.

A media-generation platform offering AI-powered tools for video creation, editing, and visual content generation, valued at $5.3 billion after a $315 million Series E round.

OpenAI's AI video generation model that creates videos from text prompts, initially released in early 2024 and considered groundbreaking at the time of its debut.

An AI company offering video and 3D generation tools, known for its Dream Machine video generation model that creates realistic video content from text and image inputs.

An AI video generation company that develops tools for creating and editing video content using artificial intelligence.

An open-source AI video generator with native sound generation capabilities. It is a mixture-of-experts model with 32 billion total parameters that can produce 360p and 720p resolution videos with synchronized audio.

An AI model by Tencent that generates videos of people interacting with objects based on text prompts, such as picking up items or holding objects, with support for stringing multiple actions together.

An AI video generation tool capable of creating highly realistic video content including avatars, animations, and creative visual sequences from text or other inputs.

An AI video generation platform by Kuaishou featuring a unique multi-shot capability that allows users to create cinematic videos with multiple customizable shots, character consistency across scenes, and support for up to 15-second videos with hard cuts.

A real-time video object removal technique from NVIDIA and collaborators that can delete objects and their secondary effects (shadows, reflections) from videos at 25 fps, using pre-trained diffusion models without additional training.

OpenAI's video generation model mentioned alongside other realistic media generation tools released in 2025.

A video generation model capable of producing realistic AI-generated video content.

xAI's video and image generation tool, reportedly generating 50 million videos a day and over 6 billion images in 30 days.

ByteDance's video generation model accessible through ByteDance's Playground Arena, requiring account creation and payment setup. Also available through ChatCard with invite codes.

An AI video generation model used as a benchmark for comparing audio synchronization and video generation quality.

Google's AI video generation model (Veo 3) capable of generating videos with native audio, referenced as a top-tier closed-source video generator.

AI company that partnered with Svedka to create what is touted as the first primarily AI-generated national Super Bowl ad, also known for AI-generated Coca-Cola commercials.