LLM Era
2020s–Present · Large language models and agentic systems
Since 2020, large language models such as GPT and Claude have developed rapidly, alongside technologies including AI agents and MCP. These advances have broadened the use of AI in content generation, software development, and tool use.
Technologies such as Claude Code, agents, and MCP allow models to edit files, run commands, and connect to external tools within granted permissions. Their capabilities and risks depend on the model, host environment, permission settings, and product version.
Representative characteristics of this period include larger models, multimodal input and output, agentic workflows, and standard protocols that connect AI to external systems. These capabilities continue to change and should be assessed against verifiable product documentation and research.
Key Milestones
GPT-3 Release
OpenAI released GPT-3, a large language model with 175 billion parameters. GPT-3 demonstrated astonishing language generation capabilities, sparking widespread discussion about AGI.
OpenAI
DALL-E & CLIP
OpenAI released DALL-E (image generation) and CLIP (image-text contrastive learning). These models demonstrated AI’s ability to understand and generate multimodal content.
OpenAI
Codex & GitHub Copilot
OpenAI released Codex, an AI model that understands and generates code. GitHub Copilot, built on Codex, became one of the first widely used AI programming assistants.
OpenAI Codex
ChatGPT Release
In November 2022, OpenAI released ChatGPT as a research preview. It brought a conversational generative-AI assistant to a mass audience and made large language models a subject of public debate.
OpenAI ChatGPT
Stable Diffusion Open-Weight Release
Stability AI released Stable Diffusion, an image generation model whose weights can be freely downloaded, under a license that restricts certain uses. Open-source community activity rapidly popularized image generation technology.
Stability AI
GPT-4 Release
OpenAI released GPT-4, a multimodal model that accepts image and text inputs. OpenAI reported human-level performance on various professional and academic benchmarks, while noting that the model remained less capable than humans in many real-world scenarios.
OpenAI
Claude Release
Anthropic released Claude, an AI assistant focused on safety and helpfulness. Claude’s design reflects emphasis on AI safety—AI should be helpful, harmless, and honest.
Anthropic Claude
LLaMA Release
Meta released LLaMA, a large language model under a non-commercial research license. Although LLaMA has relatively small parameter size, its excellent performance promoted the development of open-weight LLMs.
Meta
AI Agent Concept Rises
The AI Agent concept began to receive widespread attention. Agents are AI systems capable of autonomously planning and executing complex tasks.
AI Industry
Gemini 1.5 and Long Context
In February 2024, Google introduced Gemini 1.5 with a mixture-of-experts architecture and an experimental context window of up to one million tokens, demonstrating retrieval and reasoning across long documents, codebases, and video.
Google DeepMind
Claude 3 Series
Anthropic released the Claude 3 series, including Opus, Sonnet, and Haiku. Claude 3 Opus performed competitively with GPT-4 on several benchmarks including MMLU and GPQA, while trailing GPT-4 on others.
Anthropic
Llama 3 Open-Weight Release
Meta released the Llama 3 family with open weights, including 8B and 70B parameter models.
Meta
EU AI Act Enters into Force
On August 1, 2024 the EU AI Act entered into force. The European Commission describes it as the first comprehensive legal framework on AI worldwide. It established a risk-based classification and strict requirements for high-risk AI systems, applying in stages.
European Union
OpenAI o1 Model
OpenAI released o1-preview, a reasoning model that spends more time thinking before it answers, and reported strong results on science and programming tasks.
OpenAI
Claude Computer Use Public Beta
In October 2024, Anthropic introduced a public beta of computer use for an upgraded Claude 3.5 Sonnet. Developers could direct Claude to inspect a screen, move a cursor, click, and type, making it one of the first frontier models publicly offered with general graphical-interface control.
Anthropic Claude
MCP Protocol Release
Anthropic released MCP (Model Context Protocol), an open standard for connecting AI systems with external tools and data sources. It offers a single protocol in place of fragmented, one-off integrations.
Anthropic
Agentic Ecosystem Matures
By 2025, AI agents matured from research demos to production tools. Coding agents, browser-use agents, and multi-step task agents became standard offerings, with OpenAI, Anthropic, Google, and open-source projects all shipping capable systems.
AI Industry
DeepSeek R1 Release
Chinese AI lab DeepSeek released R1, an open-weight reasoning model, with code and weights under the MIT license, whose performance rivaled OpenAI’s o1. The release demonstrated that frontier capabilities could be achieved outside the US AI establishment and triggered a global re-evaluation of AI competition.
DeepSeek
Stargate Project Launched
OpenAI, Oracle, SoftBank, and the UAE’s MGX jointly announced that they intend to invest $500 billion over four years to build AI infrastructure in the United States, including multiple large-scale data centers.
OpenAI / Oracle / SoftBank
Claude 3.7 Sonnet Release
Anthropic released Claude 3.7 Sonnet, which it described as the first hybrid reasoning model generally available: users can switch between instant answers and extended thinking in the same model.
Anthropic
Anthropic $3.5B Funding Round
In March 2025, Anthropic announced a $3.5 billion funding round at a $61.5 billion valuation, making it one of the highest-valued AI startups in the world and reflecting further capital concentration in the AI industry.
Anthropic
Llama 4 Open-Weight Release
Meta released the natively multimodal open-weight Llama 4 Scout and Maverick models and previewed Behemoth, a teacher model that was still training and was not released. Scout and Maverick use mixture-of-experts architectures.
Meta
Claude 4 Release
Anthropic released Claude Opus 4 and Claude Sonnet 4, hybrid models that offer near-instant responses and extended thinking. Both could use tools such as web search during extended thinking, a capability then in beta, and Anthropic reported gains in coding, reasoning and long-running agentic tasks.
Anthropic
EU AI Act GPAI Obligations Begin to Apply
On 2 August 2025, the EU AI Act’s governance rules and obligations for general-purpose AI models became applicable. Other provisions follow separate transition dates, so this milestone is not a blanket start of full enforcement.
EU AI Office
GPT-5 Release
OpenAI released GPT-5, which it described as a unified system that routes between a fast model and a deeper reasoning model. OpenAI reported state-of-the-art results in mathematics, real-world coding and other benchmarks.
OpenAI
OpenAI Completes Recapitalization
OpenAI completed the recapitalization of its corporate structure. The for-profit business became a public benefit corporation named OpenAI Group PBC, and the OpenAI Foundation — the nonprofit — continues to control it under the same mission. OpenAI says the recapitalization followed nearly a year of discussions with the California and Delaware Attorneys General.
OpenAI
Agents Move Into Shipped Enterprise Products
Through the first half of 2026, agent capabilities moved out of demonstrations and into shipped enterprise products: coding, support, and office-automation agents became documented features, and tool-calling and multi-agent infrastructure matured around them. How far adoption actually reached is not settled. Published surveys report very different production rates because they do not agree on what counts as an agent, so this entry records the direction of travel rather than a level.
AI Industry
Open-Weight Models Publish at Frontier Scale
Through the first half of 2026, open-weight releases from DeepSeek, Alibaba Qwen, Moonshot, Zhipu, and Meta were published at a scale previously seen only in closed systems, with weights and model cards downloadable rather than described. Whether any of them match the closed frontier depends on which benchmark index is used and how it is weighted, so this entry records what was released and under what licence, not a ranking.
DeepSeek / Meta / Alibaba
Nature Publishes the Co-Scientist Multi-Agent System
Nature published Co-Scientist, a multi-agent system built on Gemini that formulates novel research hypotheses for experimental verification. The paper reports validation across three biomedical applications: drug repurposing, discovery of novel targets, and explaining mechanisms of antimicrobial resistance. The authors present it as an assistant to researchers rather than a replacement, which makes it peer-reviewed evidence for AI participating in hypothesis generation rather than only in analysis.
Scientific Community
Meta Begins Unwinding the Manus Acquisition
Reporting on June 11, 2026 said Meta had begun unwinding its acquisition integration with AI agent company Manus, halting data sharing and separating operations. The report linked the move to Chinese regulatory requirements; later legal and operational status should be checked against updated disclosures.
Meta / Manus
Claude Fable 5 and Mythos 5 Suspended, Then Redeployed
Anthropic launched safeguarded Claude Fable 5 for general use and the less-restricted Mythos 5 for vetted Project Glasswing partners on June 9, 2026. On June 12, U.S. export controls required restrictions on foreign nationals; unable to verify nationality in real time, Anthropic suspended both models globally. The controls were lifted on June 30. Fable 5 returned globally on July 1, while Mythos 5 initially returned for approved U.S. organizations.
Anthropic
US State Attorneys General Open Probe into OpenAI
On June 13, 2026, a coalition of US state attorneys general led by New York opened an investigation into OpenAI and issued a subpoena. The subpoena demands documents relating to advertising placement, user engagement and retention, model sycophancy, consumer and health data handling, and protections for minors and the elderly. OpenAI said it would cooperate with the attorneys general and noted that ChatGPT has built stronger safeguards for minors.
Coalition of US State Attorneys General
Samsung and SK Hynix Announce Major Multi-Year Memory Investment Plans
On 29 June 2026, Samsung and SK Hynix announced multi-year semiconductor investment plans totaling 800 trillion won and involving four memory fabs in southwestern South Korea. Dollar conversions vary with exchange rates, while scope, approvals, and execution remain subject to review.
Samsung / SK Hynix
Gemini Spark Lands on Mac as a Desktop Agent
On June 30, 2026 Google brought Gemini Spark, the agent in its Gemini app, to macOS as a beta for Google AI Ultra subscribers, starting in the US. Spark could now work across desktop files and apps, connect to more services, and track topics in real time.
Google DeepMind
Moonshot Releases Kimi K3 Open Weights
In July 2026 Moonshot AI published open weights for Kimi K3 on Hugging Face. The model card records 2.8 trillion total parameters with 104 billion activated, a 1,048,576-token context window, and the Kimi K3 License. By total parameter count it was among the largest openly downloadable language models at its release.
Moonshot AI
Cloudflare Classifies AI Crawlers by Purpose
On 1 July 2026, Cloudflare announced separate controls for Search, Agent, and Training traffic, and said that from 15 September new Cloudflare domains would block Training and Agent crawlers by default on ad-bearing pages while allowing Search by default; site owners can change the settings, and multi-purpose crawlers follow the strictest applicable rule.
Cloudflare
Microsoft Launches $2.5B AI Deployment Company
On July 2, 2026 Microsoft announced Microsoft Frontier Company, a new operating business that embeds 6,000 industry and engineering experts with customers to design and deploy AI systems, backed by a $2.5 billion Microsoft investment.
Microsoft
GPT-5.6 Becomes Generally Available After a Limited Preview
On July 9, 2026, after ending a limited preview, OpenAI made the GPT-5.6 family generally available: Sol, Terra, and Luna. During the preview, OpenAI worked closely with expert organizations and trusted partners.
OpenAI
EU AI Act Becomes Generally Applicable
On August 2, 2026 the EU AI Act became generally applicable, and the AI Office together with national authorities took up implementation, supervision, and enforcement. The AI Omnibus, in force since July 27, 2026, deferred the heaviest high-risk duties: stand-alone Annex III systems apply from December 2, 2027, and AI embedded in regulated products from August 2, 2028. Prohibited practices had already applied since February 2, 2025, and general-purpose model obligations since August 2, 2025.
OECD / G7
One Model, Two ARC-AGI-3 Scores Thirty-Six Points Apart
On September 2, 2026 ARC Prize published verified ARC-AGI-3 results for GPT-6 Astra across twelve harness configurations. On the same Semi-Private evaluation set the model scores 62.71% with the Standard harness, which lets it carry forward notes it chooses to keep, and 98.55% with a Provider Adapter harness, which preserves reasoning state between requests and compacts long conversations. Astra used fewer actions than the median tested human on 96% of levels. The figure measures a model together with its scaffolding, which is why both are recorded here.
ARC Prize Foundation
GPT-6 Astra Is Classified Critical for Cyber Capability
On September 3, 2026 OpenAI published a system card placing GPT-6 Astra at the Critical threshold for cyber capability under the company’s own Preparedness Framework, the first OpenAI model at that level. The card describes a model that can find previously unknown security flaws and develop ways to exploit them across many well-protected systems without a person guiding each step, and the sensitive capabilities are gated behind a trusted-access programme. The classification is OpenAI’s assessment of its own model against its own framework, and the card does not use the words AGI or general intelligence.
OpenAI
Key Figures
Sam Altman
OpenAI CEO
Altman leads OpenAI in advancing large language model development. ChatGPT’s release made him one of the most influential figures in the AI era. He is pushing for AGI while considering AI’s impact on society.
Dario Amodei
Anthropic CEO
Amodei was formerly VP of Research (2016–2021) at OpenAI before founding Anthropic. He emphasizes AI safety and helpfulness—Claude’s design reflects this value: AI should be helpful, harmless, and honest.
Mark Zuckerberg
Meta CEO
Zuckerberg decided to release the weights of Meta’s LLaMA models, promoting the democratization of AI technology. He believes open source benefits AI safety and development, impacting the AI ecosystem profoundly.
Ilya Sutskever
Co-founder of OpenAI, founder of Safe Superintelligence Inc.
Sutskever co-founded OpenAI and was its Chief Scientist until 2024, overseeing the research behind the GPT series. Before OpenAI he co-authored AlexNet (2012) and sequence-to-sequence learning (2014), and he was an early advocate of scaling up neural networks. In 2024 he founded Safe Superintelligence Inc., a research lab focused solely on safe superintelligent AI.
Demis Hassabis
DeepMind CEO, 2024 Nobel Laureate in Chemistry
Hassabis co-founded DeepMind and led it to breakthroughs on AlphaGo, AlphaFold, and frontier AI research. In 2024 he shared the Nobel Prize in Chemistry for AlphaFold’s contributions to protein structure prediction. He continues to argue that AI’s deepest contribution will be to scientific discovery.
Emily M. Bender
Computational Linguist
Bender is a professor of linguistics at the University of Washington and first author of On the Dangers of Stochastic Parrots, presented at FAccT in 2021. The paper argued that a language model stitches together sequences of linguistic form without reference to meaning, and it named the environmental, data, and accountability costs of building ever larger ones.
Liang Wenfeng
Founder of DeepSeek
Liang studied at Zhejiang University and applied machine learning to quantitative finance, founding the fund High-Flyer in 2015. The computing infrastructure built there later supported DeepSeek, which he founded in 2023. Its R1 model, released in 2025 under the MIT licence, was reported by its authors to perform comparably to OpenAI’s o1 on reasoning tasks, and it changed what a laboratory outside the largest Western firms was assumed to be able to publish.
Classic Quotes
We are now confident we know how to build AGI as we have traditionally understood it.
Reflections · blog.samaltman.com, 2025
I think in five years’ time it may well be able to reason better than us.
Geoffrey Hinton on the promise, risks of artificial intelligence · 60 Minutes, CBS News, 2023
I believe that open source is necessary for a positive AI future.
Open Source AI is the Path Forward · Meta, 2024
I think it could come as early as 2026, though there are also ways it could take much longer.
Machines of Loving Grace · darioamodei.com, 2024
There is potential for serious, even catastrophic, harm, either deliberate or unintentional, stemming from the most significant capabilities of these AI models.
The Bletchley Declaration by Countries Attending the AI Safety Summit, 1-2 November 2023 · GOV.UK, 2023