xAGI Labs - AI Product Studio & Research Lab

7 min read Original article ↗

November 2022 - November 2025

AI Timeline

Journey through the major events since ChatGPT revolutionized AI accessibility in November 2022

November

ChatGPT Launch

OpenAI releases ChatGPT powered by GPT-3.5 as a free research preview, marking a watershed moment in AI accessibility

December

ChatGPT Goes Viral

ChatGPT gains over 1 million users within 5 days of launch

Midjourney v4 Released

Midjourney v4 released with improved image generation

Stable Diffusion Updates

Stable Diffusion 2.0 and 2.1 launched by Stability AI

January

100 Million Users

ChatGPT surpasses 100 million users in just 2 months, becoming the fastest-growing consumer app in history

February

ChatGPT Plus Launch

OpenAI announces ChatGPT Plus subscription ($20/month) offering priority access

Bing AI Integration

Microsoft announces ChatGPT-powered features coming to Bing

Bing Chat Mobile Preview

Microsoft releases AI-powered Bing Chat for mobile preview

Meta Releases LLaMA

Meta releases LLaMA language model as open-source for research

March

ChatGPT API Launch

OpenAI introduces ChatGPT API for developers; early adopters include Snapchat, Quizlet, Instacart

GPT-4 Release

OpenAI releases GPT-4 with multimodal capabilities (text and image inputs), marking major advancement in AI capabilities

Baidu Ernie Bot

Baidu releases Ernie Bot, China's ChatGPT competitor

Google Bard Launch

Google opens access to Bard chatbot based on LaMDA

AI Development Pause Letter

Open letter signed by tech leaders calling for 6-month pause on advanced AI training

Midjourney v5 Launch

Midjourney v5 launched with significant improvements

April

Data Controls Added

OpenAI adds data controls allowing users to opt out of training data inclusion

Italy Lifts ChatGPT Ban

Italy lifts ChatGPT ban after OpenAI meets privacy demands

ChatGPT Plugins Launch

OpenAI releases ChatGPT plugins, GPT-3.5 with browsing, and GPT-4 with browsing (ALPHA)

Adobe Firefly Beta

Adobe releases Firefly image creation model (beta)

May

200+ Plugins Available

ChatGPT Plus users gain access to over 200 plugins

ChatGPT iOS App

OpenAI launches ChatGPT iOS app worldwide

Senate Testimony

OpenAI CEO Sam Altman testifies before Senate on AI regulation

Bing Powers ChatGPT

Microsoft announces Bing will power ChatGPT web browsing

Midjourney v5.1

Midjourney v5.1 released

Bard Upgraded to PaLM 2

Google announces upgrade to Bard with PaLM 2, expanding to 180 countries

June

Midjourney v5.2

Midjourney v5.2 launched with enhanced features

July

ChatGPT Android App

OpenAI launches official ChatGPT app for Android

Claude 2 Released

Anthropic releases Claude 2 with 100,000 token context window (75,000 words)

LLaMA 2 Public Release

Meta releases LLaMA 2 open-source to general public in various sizes

Stable Diffusion XL 1.0

Stable Diffusion XL 1.0 released

August

ChatGPT Enterprise

OpenAI launches ChatGPT Enterprise with unlimited GPT-4 access and enterprise-level security

September

DALL-E 3 Integration

ChatGPT integrates with DALL-E 3

Voice & Image Capabilities

Voice and image capabilities launched for ChatGPT mobile apps

October

DALL-E 3 Released

DALL-E 3 released with improved image generation

Adobe Firefly 2

Adobe releases Firefly 2

November

Custom GPTs Announced

OpenAI announces Custom GPTs, allowing users to create specialized GPT versions

Sam Altman Crisis

Sam Altman fired and then reinstated as OpenAI CEO after dramatic board crisis

GPT-4 Turbo

GPT-4 Turbo announced with 128K context window and knowledge cutoff extended to April 2023

December

Stable Diffusion XL Turbo

Stable Diffusion XL Turbo released (one-step real-time generation)

Midjourney v6

Midjourney v6 launched

Gemini Pro in Bard

Google upgrades Bard to Gemini Pro in limited areas

Grok AI Launch

X Corporation launches Grok AI chatbot for paid subscribers

January

GPT Store Launch

OpenAI launches GPT Store for discovering custom GPTs and introduces ChatGPT Team

Stable LM 2

Stability AI releases Stable LM 2 with up to 12 billion parameters

New Embedding Models

OpenAI releases text-embedding-3-small and text-embedding-3-large models

Stable Diffusion 3 Announced

Stability AI announces Stable Diffusion 3 (gradual rollout to waitlist)

February

Bard Becomes Gemini

Google's Bard officially rebrands to Gemini

Sora Announcement

OpenAI announces Sora, groundbreaking text-to-video model

Model Context Protocol

Anthropic announces Model Context Protocol (MCP) - a universal interface for AI tools

Gemini 1.0 Ultra

Google launches Gemini 1.0 Ultra for select developers

March

Claude 3 Family

Anthropic releases Claude 3 family (Haiku, Sonnet, Opus) with meta-cognitive reasoning

Inflection-2.5

Inflection AI releases Inflection-2.5 achieving 94% of GPT-4 performance with 40% of training compute

Jamba Released

AI21 Labs unveils Jamba with 256K context window

Stable Audio 2.0

Stability AI releases Stable Audio 2.0

Grok-1.5V

X Corporation releases Grok-1.5V with image recognition

April

Plugins Discontinued

OpenAI announces discontinuation of ChatGPT plugins in favor of custom GPTs

Mixtral 8x22B

Mistral AI launches Mixtral 8x22B (141B total, 39B active parameters)

LLaMA 3 Released

Meta releases LLaMA 3 in 8B and 70B sizes (open-source)

Microsoft Phi-3-mini

Microsoft releases Phi-3-mini (3.8B parameters, mobile-capable)

Adobe Firefly 3

Adobe announces Firefly 3 for image creation

May

GPT-4o Launch

OpenAI announces GPT-4o (omni) with native multimodal capabilities and faster processing

Google I/O AI Showcase

Google I/O showcases major AI advancements

Claude 3.5 Sonnet

Anthropic releases Claude 3.5 Sonnet with major improvements in coding and reasoning

Meta Chameleon

Meta introduces Chameleon multimodal model

Google AI Overviews

Google announces AI Overviews for search

Apple Intelligence

Apple announces Apple Intelligence for its devices

June

Stable Diffusion 3 Medium

Stability AI releases Stable Diffusion 3 medium version

Runway Gen3 Alpha

Runway introduces Gen3 Alpha for video generation

Gemma 2

Google announces Gemma 2

Gemini 1.5 Upgrade

Google increases Gemini 1.5 token limit to 2 million

Copilot+ PCs

Microsoft announces Copilot+ for dedicated AI PCs

July

GPT-4o mini

OpenAI releases GPT-4o mini

Llama 3.1

Meta releases Llama 3.1 in multiple sizes including 405B parameters

Mistral Large 2

Mistral AI releases Mistral Large 2 (123B parameters)

AlphaProof & AlphaGeometry 2

Google DeepMind unveils AlphaProof and AlphaGeometry 2

SearchGPT Launch

OpenAI launches SearchGPT

Midjourney v6.1

Midjourney v6.1 released

August

Advanced Voice Mode

Advanced Voice Mode rolled out for ChatGPT with natural real-time conversations

Flux 1.1 Pro

Flux 1.1 Pro released

September

Meta Movie Gen

Meta unveils Movie Gen

Pika Video 1.5

Pika introduces Video Model 1.5 with Pika Effects

October

Adobe Firefly Video

Adobe announces Firefly Video

Stable Diffusion 3.5

Stable Diffusion 3.5 released

Claude 3.5 Updates

Anthropic launches Claude 3.5 Sonnet (new version) and Claude 3.5 Haiku

Recraft v3

Recraft v3 released

November

MCP Production Ready

Anthropic introduces Model Context Protocol (MCP) as production-ready

Pixtral Large

Mistral AI launches Pixtral Large

Visual PDF Analysis

Anthropic introduces Visual PDF Analysis in Claude

December

o3 Preview

OpenAI previews o3 reasoning model with major performance improvements

Sora Public Release

OpenAI releases Sora for video generation (public access)

o1 Reasoning Model

OpenAI releases o1 reasoning model

ChatGPT Pro

ChatGPT Pro tier introduced at $200/month with unlimited o1 access

Gemini 2.0 Flash

Google launches Gemini 2.0 Flash in beta

Gemini Thinking Mode

Google introduces Gemini-2.0-Flash-Thinking

Veo 2

Google releases Veo 2 for video generation

January

Trump Inauguration

Donald Trump inaugurated as US President

DeepSeek R1 Disruption

DeepSeek releases R1 reasoning model, disrupting AI industry with claimed $6M training cost and performance matching OpenAI o1

o3-mini for Free Users

OpenAI releases o3-mini reasoning model to free ChatGPT users

February

o3-mini Enhanced Reasoning

OpenAI updates o3-mini to show more detailed chain-of-thought reasoning in response to DeepSeek competition

GPT-4.5 for Pro Users

GPT-4.5 (codename Orion) rolled out to ChatGPT Pro users

Alibaba Qwen Updates

Alibaba releases Qwen updates claiming to beat DeepSeek

May

Gemini 1.5 Milestone

Google releases Gemini 1.5 (often called 2.5), first to break 1400 Elo rating on Arena LLM Leaderboard for web development

Claude 4 Family

Anthropic introduces Claude 4 family (Opus 4 and Sonnet 4) with 200K token context and extended thinking mode

Claude Code GA

Claude Code becomes generally available with GitHub Actions support

Gemini 2.5 Multi-Agent

Google releases Gemini 2.5 Pro and Flash as first publicly available multi-agent model with 1M token context

June

Claude 4 GA

Claude 4 becomes generally available via Claude.ai and API

AI Safety Level 3

Anthropic announces Claude 4 operates at AI Safety Level 3

August

OpenAI Open-Source Models

OpenAI launches gpt-oss-120b and gpt-oss-20b, its first open-source reasoning models, responding to Chinese AI lab pressure

September

700M Weekly Users

ChatGPT reaches 700 million weekly active users

80% Market Share

OpenAI reportedly accounts for 80% of generative AI tool traffic

November

Gemini 3 Launch

Google launches Gemini 3 with advanced multimodal, agentic capabilities and 'Deep Think' mode

Claude on Azure

Microsoft announces Anthropic's Claude models available in Microsoft Foundry (Azure)

Midjourney v7

Midjourney launches v7 with higher image quality and style control

Video Generation Advances

Veo 2.0 (Google), Runway Gen-4, Vidu Q1, and Kling 2.0 released

Qwen 3

Alibaba releases Qwen 3

Phi-4 Reasoning

Microsoft launches Phi-4 reasoning series

Veo 3 & Imagen 4

Google releases Veo 3 for 4K video with natural audio and Imagen 4

Key Highlights

700M+

ChatGPT Weekly Active Users (2025)

3+ Years

Since ChatGPT Launch

50+

Major AI Model Releases

Key Themes & Trends

Model Evolution

From foundation models (2022-2023) to multimodal capabilities (2024) to reasoning models (2025)

Open Source Movement

Meta's LLaMA series, Mistral AI, and DeepSeek drive open-source innovation and cost efficiency

Multimodal Expansion

From text-only to images, voice, video, and seamless integration across modalities

Competition Intensifies

US-China rivalry, open vs closed debate, and rapid innovation cycles reshape the industry