Gemini 4 Neon
Google introduces Gemini 4 Neon for software engineering, knowledge and science work.
AI Lists βΊ AI Developments
A dated timeline of the AI releases, laws and milestones that matter, newest first.
Each development carries the date from its official source. Newest entries are at the top of each section.
Google introduces Gemini 4 Neon for software engineering, knowledge and science work.
Google's fast Gemini 4 model for everyday coding and agentic workflows.
Anthropic launches Fable 5 for general use and Mythos 5 for vetted cyber defenders.
Google starts the Gemini 3.5 family with Flash, aimed at agents and coding.
OpenAI introduces GPT-6.2 Astra (frontier) and Sol (efficient everyday model).
OpenAI releases GPT-5 in ChatGPT and the API.
OpenAI's first open-weight language models since GPT-2.
xAI releases Grok 4.
Anthropic's Claude 4 models, focused on coding and long-running agents.
Google announces an open protocol for agents to work together.
Meta releases Llama 4 Scout and Maverick, natively multimodal open-weight models.
Google's 'thinking' model family debuts with Gemini 2.5 Pro.
First hybrid reasoning Claude model, plus a preview of Claude Code.
OpenAI and partners announce a large US AI infrastructure venture.
Open-weight reasoning model from DeepSeek released under the MIT license.
Anthropic open-sources MCP for connecting AI to tools and data.
Claude 3.5 Sonnet gains a public beta for operating a computer.
Awarded to David Baker, Demis Hassabis and John Jumper.
Awarded to John Hopfield and Geoffrey Hinton.
o1-preview introduces models that think before answering.
OpenAI's natively multimodal model for text, audio and vision.
Anthropic releases Claude 3 Haiku, Sonnet and Opus.
OpenAI previews its text-to-video model.
Google DeepMind introduces Gemini.
Meta releases Llama 2 with a license allowing commercial use.
OpenAI's GPT-4 accepts image and text input.
OpenAI opens ChatGPT to the public as a research preview.
Open image model released to the public.
OpenAI's image model makes realistic pictures from text.
'Language Models are Few-Shot Learners' shows scaling unlocks in-context learning.
The paper that introduced the Transformer architecture.
DeepMind's AlphaGo wins 4-1 against a top Go professional.
The world's first broad AI law starts its phased rollout.
Voluntary US framework for managing AI risk (AI RMF 1.0).
For AI agents and tools: JSON Β· Markdown Β· llms.txt Β· RSS. Page built Nov 20, 2026, 12:17 AM PT.