Timeline of AI Models Cheat Sheet

Leave a comment

Below is a timeline of all AI models from some of the major vendors. Note, data may be incomplete or inaccurate. Assume up-to the publishing date.

DateVendorModel / familyTypeStrengths / expertise
Jun-18🟢 OpenAIGPT-1🧠 General · 🔬 ResearchThe foundational Generative Pre-trained Transformer: demonstrated transfer learning from large-scale language pretraining.
Feb-19🟢 OpenAIGPT-2🧠 GeneralMajor leap in generative text quality · ✍️ Writing · 💬 Conversation. Initially released cautiously because of misuse concerns.
May-20🟢 OpenAIGPT-3🧠 General🚀 Major scaling milestone · ✍️ Writing · 💻 Code generation · few-shot learning.
Nov-20🟢 OpenAIDALL·E🎨 Image · 🔬 ResearchEarly major text-to-image model demonstrating image generation from natural-language prompts.
Jan-21🟢 OpenAICodex💻 CodingNatural-language-to-code generation; strategically important precursor to modern AI coding assistants.
Jan-21🟢 OpenAIDALL·E public research announcement🎨 ImageText-to-image generation became a major strategic model category.
Jun-21🟢 OpenAICodex API / Codex family💻 Coding💻 Software generation · natural-language programming · foundation for GitHub Copilot-era workflows.
Apr-22🟢 OpenAIDALL·E 2🎨 ImageMajor improvement in image realism, editing and variation generation.
Nov-22🟢 OpenAIWhisper🎙️ Voice · 🔓 Open-weightStrategically important speech-recognition family · transcription · multilingual speech recognition.
Nov-22🟢 OpenAIGPT-3.5 / ChatGPT generation🧠 General · 💬 Conversation💬 Conversational AI breakthrough that brought LLMs into mainstream consumer use.
Nov-22🟦 MicrosoftVALL-E🎙️ Voice · 🔬 ResearchImportant neural text-to-speech and voice-generation research model.
Mar-23🟢 OpenAIGPT-4🌐 Multimodal · 🧠 GeneralMajor frontier-model leap · 🧠 reasoning · 💻 coding · 👁️ vision capability in multimodal versions.
Mar-23🟦 MicrosoftKosmos-1🌐 Multimodal · 🔬 ResearchMicrosoft’s early multimodal language-model research combining language and perception.
Mar-23🟠 AnthropicClaude 1🧠 General · 💬 ConversationAnthropic’s first major public Claude generation; emphasis on helpfulness and safety.
Apr-23🟢 OpenAIGPT-3.5 Turbo🧠 General · ⚡ EfficientFaster, lower-cost production/API-oriented GPT-3.5 generation.
Apr-23🟧 Amazon / AWSAmazon Titan foundation models💼 Professional/enterpriseAWS’s strategically important in-house foundation-model family for Bedrock and enterprise applications.
Jul-23🟠 AnthropicClaude 2🧠 General · 📚 Long-contextMajor context-window expansion · writing · analysis · enterprise use.
Jul-23🔷 MetaLlama 2🔓 Open-weight · 🧠 GeneralMajor open-model ecosystem milestone · 🧩 customization · 🏠 self-hosting.
Sep-23🟢 OpenAIDALL·E 3🎨 ImageMuch stronger prompt following and integration with conversational generation.
Sep-23🟠 AnthropicClaude 2.1🧠 General · 📚 Long-contextImproved reliability and a strategically important 200K-token context window.
Sep-23🟦 MicrosoftPhi-1 / Phi-1.5 family⚡ Small/efficientDemonstrated Microsoft’s strategy around surprisingly capable small language models.
Nov-23🟧 Amazon / AWSTitan Image Generator🎨 ImageAWS-native image-generation foundation model for Bedrock.
Nov-23🟧 Amazon / AWSTitan Multimodal Embeddings🌐 Multimodal · 💼 EnterpriseMultimodal retrieval and enterprise search/RAG applications. (Amazon Web Services, Inc.)
Dec-23🔵 Google / Google DeepMindGemini 1.0🌐 Multimodal · 🧠 GeneralGoogle’s new flagship multimodal family, including Ultra, Pro and Nano tiers.
Dec-23🔵 Google / Google DeepMindGemini Nano⚡ Small/efficient · 📱 On-deviceStrategically important efficient/on-device AI direction.
Dec-23🟦 MicrosoftPhi-2⚡ Small/efficientStrong small-model capability for its size; important for local and constrained deployment.
Feb-24🔵 Google / Google DeepMindGemma🔓 Open-weightGoogle’s major open-model family, optimized for customization and developer deployment.
Feb-24🟢 OpenAISora🎥 VideoLandmark text-to-video model announcement; demonstrated long-form generative video capabilities.
Feb-24🟢 OpenAIGPT-4 Turbo🧠 General · 📚 Long-contextFaster/cheaper GPT-4 generation with expanded context and stronger production usability.
Mar-24🟠 AnthropicClaude 3 Haiku⚡ Small/efficientSpeed and affordability for high-volume workloads.
Mar-24🟠 AnthropicClaude 3 Sonnet🧠 General · 💼 ProfessionalBalanced capability, speed and cost for professional work.
Mar-24🟠 AnthropicClaude 3 Opus🏆 Frontier/high-endAnthropic’s highest-capability Claude 3 model · reasoning · writing · analysis.
Apr-24🔷 MetaLlama 3🔓 Open-weight · 🧠 GeneralMajor open-weight release that significantly strengthened the Llama ecosystem.
Apr-24🟦 MicrosoftPhi-3 family⚡ Small/efficient · 🔓 Open-weightSmall models for edge, local and enterprise deployment; important strategic expansion of Phi.
May-24🟧 Amazon / AWSTitan Text Premier💼 Professional/enterprise · 📚 RAGAmazon’s most advanced Titan text model, optimized for enterprise RAG and agent applications. (Amazon Web Services, Inc.)
May-24🟢 OpenAIGPT-4o🌐 Multimodal · 🎙️ VoiceMajor “omni” model milestone: text, vision and near-real-time audio interaction in one family.
Jun-24🟠 AnthropicClaude 3.5 Sonnet🧠 General · 💻 CodingMajor jump in 💻 coding, vision and professional reasoning; became highly influential for agentic coding.
Jun-24🔵 Google / Google DeepMindGemini 1.5 Pro🌐 Multimodal · 📚 Long-contextMajor long-context milestone, including the 1M-token context direction.
Jun-24🔵 Google / Google DeepMindGemini 1.5 Flash⚡ Small/efficient · 🌐 MultimodalFaster, cheaper multimodal model for scaled applications.
Jul-24🔷 MetaLlama 3.1🔓 Open-weight · 🧠 GeneralExpanded open-weight frontier with major larger-scale models and enterprise/self-hosting momentum.
Jul-24🟢 OpenAIGPT-4o mini⚡ Small/efficient · 🌐 MultimodalCost-efficient model designed to bring capable multimodal AI to high-volume applications.
Aug-24🟢 OpenAIGPT-4.5 preview-era research trajectory🧠 GeneralNot counted here as a 2024 public GPT-4.5 release; the verified GPT-4.5 public release belongs later.
Sep-24🟢 OpenAIo1-preview🧠 ReasoningOpenAI’s major public reasoning-model transition: extended inference for harder math, science and coding problems.
Sep-24🟢 OpenAIo1-mini🧠 Reasoning · 💻 CodingMore efficient reasoning model, especially important for STEM and coding workloads.
Sep-24🔷 MetaLlama 3.2🔓 Open-weight · 🌐 Multimodal · 📱 EdgeAdded strategically important vision models and smaller edge/mobile-oriented models.
Oct-24🟠 AnthropicClaude 3.5 Haiku⚡ Small/efficient · 💻 CodingFaster, lower-cost Claude optimized for production-scale workloads.
Oct-24🟠 AnthropicClaude 3.5 Sonnet (new)💻 Coding · 🤖 AgenticImportant upgrade in coding, computer interaction and agent-style workflows.
Oct-24🟢 OpenAIGPT-Realtime / gpt-realtime🎙️ Voice · 🌐 MultimodalNative real-time speech interaction; strategically important for low-latency voice agents.
Dec-24🟢 OpenAIo1🧠 ReasoningProduction reasoning model succeeding the preview generation; 🧠 reasoning · ➗ mathematics · 💻 coding.
Dec-24🟧 Amazon / AWSAmazon Nova family🌐 Multimodal · 💼 Professional/enterpriseNova Micro, Lite, Pro and Canvas/Reel families established AWS’s new broad foundation-model strategy. (Amazon Web Services, Inc.)
Dec-24🔴 DeepSeekDeepSeek-V3🔓 Open-weight · 🧠 General · 💰 Cost efficiencyLarge MoE open model emphasizing strong capability, speed and cost efficiency. (DeepSeek)
Jan-25🔴 DeepSeekDeepSeek-R1🧠 Reasoning · 🔓 Open-weightMajor global reasoning/open-model milestone · 🧠 reasoning · ➗ mathematics · 💻 coding. Released with MIT licensing and distilled variants. (DeepSeek)
Feb-25🟢 OpenAIo3-mini🧠 Reasoning · 💻 CodingMore accessible and efficient advanced reasoning.
Feb-25🟠 AnthropicClaude 3.7 Sonnet🧠 Reasoning · 💻 Coding · 🤖 AgenticHybrid reasoning approach with extended thinking; major step toward modern coding agents.
Feb-25🟢 OpenAIGPT-4.5🧠 General · ✍️ WritingLarge-scale general-purpose GPT release emphasizing natural interaction and knowledge work.
Mar-25🔵 Google / Google DeepMindGemma 3🔓 Open-weight · 🌐 MultimodalGoogle’s next major open-model generation with stronger multimodal capabilities.
Mar-25🔵 Google / Google DeepMindGemini 2.5 Pro🧠 Reasoning · 🌐 MultimodalMajor reasoning-focused Gemini generation with “thinking” capabilities.
Apr-25🟢 OpenAIo3🧠 Reasoning · 🤖 AgenticHigh-end reasoning model with strong tool use, coding and complex multi-step problem solving.
Apr-25🟢 OpenAIo4-mini🧠 Reasoning · 💻 CodingEfficient reasoning model with strong STEM and tool-use capability.
Apr-25🟧 Amazon / AWSAmazon Nova Premier🏆 Frontier/high-end · 🌐 Multimodal · 📚 Long-contextAWS’s most capable Nova model for complex planning, tool use and million-token-scale context. (Amazon Web Services, Inc.)
Apr-25🔷 MetaLlama 4🔓 Open-weight · 🌐 MultimodalMajor new Llama generation emphasizing multimodality and MoE-scale architectures.
May-25🟠 AnthropicClaude Sonnet 4💻 Coding · 🤖 AgenticStrong software engineering and agentic task execution.
May-25🟠 AnthropicClaude Opus 4🏆 Frontier/high-end · 💻 Coding · 🤖 AgenticHigh-end agentic coding and long-running task execution.
Jun-25🟢 OpenAIGPT-Image-1 / GPT Image family🎨 ImageOpenAI’s strategically important modern image-generation API family.
Jun-25🟢 OpenAICodex agent generation💻 Coding · 🤖 AgenticModern coding-agent evolution: autonomous software tasks rather than simply code completion.
Aug-25🟠 AnthropicClaude Opus 4.1💻 Coding · 🧠 ReasoningIncremental but strategically significant Opus improvement for advanced coding and agentic work.
Aug-25🔴 DeepSeekDeepSeek-V3.1🤖 Agentic · 🧠 Reasoning · 🔓 Open-weightHybrid Think/Non-Think inference and stronger tool use; DeepSeek explicitly positioned it as a step toward the agent era. (DeepSeek)
Sep-25🟠 AnthropicClaude Sonnet 4.5💻 Coding · 🤖 AgenticMajor Sonnet-class upgrade for autonomous coding and professional workflows.
Sep-25🔴 DeepSeekDeepSeek-V3.2-Exp🔓 Open-weight · 📚 Long-contextExperimental DeepSeek Sparse Attention architecture targeting more efficient long-context inference.
Oct-25🟠 AnthropicClaude Haiku 4.5⚡ Small/efficient · 💻 CodingFast and economical modern Claude generation.
Nov-25🟠 AnthropicClaude Opus 4.5🏆 Frontier/high-end · 💻 Coding · 🤖 AgenticStrong frontier agentic and coding capability.
Nov-25🔵 Google / Google DeepMindGemini 3 / Gemini 3 Pro🧠 Reasoning · 🌐 Multimodal · 🤖 AgenticMajor new Gemini generation combining advanced reasoning, multimodality and agentic capability. (blog.google)
Dec-25🔵 Google / Google DeepMindGemini 3 Flash⚡ Small/efficient · 🌐 Multimodal · 🧠 ReasoningHigh-speed model combining strong reasoning with image, audio and video understanding. (blog.google)
Dec-25🔴 DeepSeekDeepSeek-V3.2🧠 Reasoning · 🤖 Agentic · 🔓 Open-weightReasoning-first successor with integrated thinking during tool use. (DeepSeek)
Dec-25🔴 DeepSeekDeepSeek-V3.2-Speciale🧠 Reasoning · ➗ Mathematics · 🔬 SpecializedHigh-end reasoning variant focused on extremely difficult mathematics and competition-style reasoning. (DeepSeek)
Dec-25🟧 Amazon / AWSNova 2 Sonic🎙️ Voice · 🌐 MultimodalGeneral-availability speech-to-speech foundation model for real-time conversational AI. (Amazon Web Services, Inc.)
5-Feb-26🟠 AnthropicClaude Opus 4.6🏆 Frontier/high-end · 💻 Coding · 🤖 AgenticBetter planning, longer-running agentic tasks, codebase work and a 1M-token context beta. (Anthropic)
########🟠 AnthropicClaude Sonnet 4.6💻 Coding · 🤖 Agentic · 📚 Long-contextMajor Sonnet upgrade across coding, computer use, planning and knowledge work. (Anthropic)
Apr-26🔴 DeepSeekDeepSeek-V4 Preview🔓 Open-weight · 📚 Long-contextOfficially released/open-sourced preview emphasizing cost-effective million-token context. (DeepSeek)
########🟠 AnthropicClaude Opus 4.8🏆 Frontier/high-end · 💻 Coding · 🤖 AgenticImproved coding, agentic skills and practical knowledge work; added more flexible effort/speed controls. (Anthropic)
########🔵 Google / Google DeepMindGemini 3.5 Flash🤖 Agentic · 💻 Coding · ⚡ EfficientMajor transition toward frontier intelligence with action, optimized for long-horizon agentic workflows. (blog.google)
########🔵 Google / Google DeepMindGemini Omni Flash🌐 Multimodal · 🎥 VideoNative multimodal generation from mixed inputs, including video-oriented creation and editing. (blog.google)
24-Jun-26🔵 Google / Google DeepMindGemini 3.5 Flash Computer Use🖥️ Computer use · 🤖 AgenticComputer interaction became integrated into the main Flash model for browser, mobile and desktop agents. (blog.google)
30-Jun-26🟠 AnthropicClaude Sonnet 5🤖 Agentic · 💻 Coding · 🧠 ReasoningExplicitly designed as Anthropic’s most agentic Sonnet model, with planning, browser and terminal tool use. (Anthropic)
9-Jul-26🟢 OpenAIGPT-5.6 family🧠 General · 💼 Professional · 🤖 AgenticSol, Terra and Luna tiers: stronger efficiency, coding, science, cybersecurity and multi-agent work. (OpenAI)
21-Jul-26🔵 Google / Google DeepMindGemini 3.6 Flash⚡ Efficient · 💻 Coding · 🤖 AgenticNew Flash workhorse focused on token efficiency, lower latency and production agents. (blog.google)
21-Jul-26🔵 Google / Google DeepMindGemini 3.5 Flash-Lite⚡ Small/efficientLower-cost/high-throughput Gemini generation for scaled workloads. (blog.google)
24-Jul-26🟠 AnthropicClaude Opus 5🏆 Frontier/high-end · 💻 Coding · 💼 ProfessionalNew flagship everyday Opus-class model with state-of-the-art coding and knowledge-work positioning. (Anthropic)
3-Sep-26🟢 OpenAIGPT-6 Astra🧠 Reasoning · 🖥️ Computer use · 🤖 Agentic · 🏆 FrontierOpenAI’s newest major verified model as of the cutoff: advanced computer use, browsing, software engineering, cybersecurity, science and professional work. (OpenAI)
########🔴 DeepSeekDeepSeek-V4.1-Flash⚡ Efficient · 👁️ Vision · 🤖 Agentic · 🔓 Open-weightNew architecture family with native visual understanding, asymmetric MoE design and efficiency aimed at high-throughput agentic workloads. (DeepSeek)

Specialized AI Model Families

VendorFamilyPrimary specialtyWhat it’s good for
🟢 OpenAIWhisper🎙️ Speech recognitionMultilingual transcription and speech translation
🟢 OpenAIDALL·E🎨 Image generationText-to-image creation and editing
🟢 OpenAIGPT-image🎨 ImageNative multimodal image creation/editing and text rendering O OpenAI
🟢 OpenAISora🎥 VideoGenerative video and video transformation
🟢 OpenAICodex💻 Coding agentsRepository-scale development and autonomous software engineering
🟢 OpenAIGPT-Live🎙️ VoiceFull-duplex real-time conversational AI O OpenAI
🟢 OpenAIGPT-6 Astra🖥️ Computer-use agentsComputer operation, browsing, professional workflows, coding and research O OpenAI
🟠 AnthropicClaude Computer Use🖥️ Computer useBrowser/desktop interaction and UI automation
🟠 AnthropicClaude Code💻 Coding agentsLong-running software engineering
🟠 AnthropicClaude Agent SDK🤖 Agent framework/model toolingBuilding autonomous agents with memory, permissions and tool use
🔵 GoogleGemma🔓 Open-weight / smallLocal inference, customization and edge deployment
🔵 GoogleVeo🎥 VideoGenerative video
🔵 GoogleGemini Image🎨 ImageImage generation/editing
🔵 GoogleGemini Audio / Live🎙️ AudioReal-time voice and audio reasoning
🔵 GoogleGemini Computer Use🖥️ Computer useOperating browser/software interfaces
🔵 GoogleGemini 3.5 agentic family🤖 AgentsLong-horizon action and tool execution B blog.google
🔷 MetaLlama Vision👁️ VisionImage understanding in customizable/open deployments M Meta AI
🔷 MetaCode Llama💻 CodingCode generation and completion
🔷 MetaLlama Guard🛡️ SafetyContent moderation and safety classification
🔴 DeepSeekDeepSeek-R1🧠 ReasoningMath, science, coding and open-weight reasoning
🔴 DeepSeekDeepSeek-V3 family🧠 General / 💻 CodingEfficient general intelligence and coding
🔴 DeepSeekDeepSeek-V4🌐 Multimodal / 🤖 AgentsLong-context reasoning, vision and agentic work D DeepSeek API Docs
🟦 MicrosoftPhi⚡ Small modelsEfficient reasoning, edge and local deployment
🟦 MicrosoftPhi reasoning🧠 Small reasoningMath, science and coding on compact models
🟧 AmazonTitan💼 Enterprise foundation modelsText, embeddings, image generation and Bedrock applications
🟧 AmazonNova🌐 MultimodalEnterprise text, image/video and agent workloads A Amazon Web Services, Inc.
🟧 AmazonNova 2🧠 Reasoning / 🤖 AgentsExtended thinking, tools and 1M-token enterprise workflows A Amazon Web Services, Inc.
🟧 AmazonNova Sonic🎙️ VoiceReal-time speech-to-speech applications A Amazon Web Services, Inc.

What Each Vendor is Becoming Known By

Vendor🏆 Core identityParticularly strong at
🟢 OpenAIFrontier general intelligence increasingly designed as an agentic computer operator🧠 Reasoning · 💻 Coding · 🤖 Agents · 🖥️ Computer use · 🎙️ Voice · 🎨 Image
🟠 AnthropicCoding and long-running professional agents💻 Coding · 🤖 Agents · 🖥️ Computer use · 📚 Long context · 💼 Knowledge work
🔵 Google DeepMindNative multimodality + agents + full-stack AI🌐 Multimodal · 🤖 Agents · 💻 Coding · 🎙️ Audio · 🎨 Image · 🎥 Video
🔷 MetaOpen-weight/customizable AI ecosystem🔓 Open weights · 🧩 Customization · 📱 Edge · 🌐 Multimodality
🔴 DeepSeekEfficient open-weight reasoning🧠 Reasoning · ➗ Math · 💻 Coding · 💰 Efficiency · 🔓 Open weights
🟦 MicrosoftSmall efficient models for local/enterprise computing⚡ Small models · 📱 Edge · 🧠 Efficient reasoning · 💼 Enterprise
🟧 Amazon/AWSCloud-delivered foundation models and enterprise agents☁️ Bedrock · 💼 Enterprise · 💰 Price/performance · 🎙️ Voice · 🤖 Agents

 

THANKS FOR READING. BEFORE YOU LEAVE, I NEED YOUR HELP.
 

I AM SPENDING MORE TIME THESE DAYS CREATING YOUTUBE VIDEOS TO HELP PEOPLE LEARN THE MICROSOFT POWER PLATFORM.

IF YOU WOULD LIKE TO SEE HOW I BUILD APPS, OR FIND SOMETHING USEFUL READING MY BLOG, I WOULD REALLY APPRECIATE YOU SUBSCRIBING TO MY YOUTUBE CHANNEL.

THANK YOU, AND LET'S KEEP LEARNING TOGETHER.

CARL

https://www.youtube.com/carldesouza

 

ABOUT CARL DE SOUZA

Carl de Souza is a developer and architect focusing on Microsoft Dynamics 365, Power BI, Azure, and AI.

carldesouza.comLinkedIn Twitter | YouTube

 

Leave a Reply

Your email address will not be published. Required fields are marked *