https://agi-milestones.pages.dev/en/ AI milestones. Item dates are event dates; revision history is available on the website. en Fri, 09 Oct 2026 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-openai-open-math-research-collection-2026https://agi-milestones.pages.dev/en/#event-openai-open-math-research-collection-2026OpenAI released mathematical research produced by an unreleased internal model: the initial 722 manuscripts were grouped into 372 result families from an evaluation of roughly 4,000 open research problems. Most results used the same procedure, averaging compute equivalent to about three hours of ChatGPT Pro thinking per result. The repository includes Lean formalizations of selected results and checking configurations. Verification varies, and some manuscripts have been withdrawn or revised; manuscript counts are not counts of confirmed solutions to distinct problems.Tue, 06 Oct 2026 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-gemini-4-argon-trusted-rollout-2026https://agi-milestones.pages.dev/en/#event-gemini-4-argon-trusted-rollout-2026Google began rolling out its new frontier model Gemini 4 Argon through the Fairwind Program, raising the output limit from 64K to 1M tokens for long-horizon coding, professional work and cyber defense. Google reports that its agents applied memory optimizations across its data centers, freeing more than 300 TiB of memory. These effects are reported by Google; the model is not yet broadly available to developers and consumers.Wed, 30 Sep 2026 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-openai-dots-2026https://agi-milestones.pages.dev/en/#event-openai-dots-2026OpenAI began rolling out GPT-6 Astra-powered dots to eligible Pro and Business Premium users. Each dot has its own cloud computer, can use authorized connected apps, continue tasks across conversations, and look for ways to help in the background. Permissions and action reviews govern steps that affect accounts or share information.Tue, 29 Sep 2026 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-glm-53-open-cyber-capability-verified-2026https://agi-milestones.pages.dev/en/#event-glm-53-open-cyber-capability-verified-2026Anthropic evaluated Z.ai's released open-weight GLM-5.3 in isolated environments. It completed end-to-end exploits in 50 of 410 attempts on a browser vulnerability benchmark and, with researcher guidance, found and chained previously unknown browser flaws into a page able to read local files. The US NIST had earlier rated it the most cyber-capable open-weight model then released. These controlled results do not establish that a real-world attack occurred.Tue, 29 Sep 2026 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-openai-navier-stokes-solution-2026https://agi-milestones.pages.dev/en/#event-openai-navier-stokes-solution-2026OpenAI says an unreleased internal model significantly more capable than GPT-6 Astra powered roughly 10,000 concurrent agents to construct, in about 88 hours, an analytical proof that a smooth, forced three-dimensional incompressible Navier–Stokes flow can develop a finite-time singularity from rest. GPT-6 Astra then completed a Lean formalization and verification in another 17 hours. OpenAI released the 166-page paper and proof repository; the result still awaits independent mathematical review.Tue, 08 Sep 2026 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-claude-fermat-formalization-2026https://agi-milestones.pages.dev/en/#event-claude-fermat-formalization-2026Dozens of Anthropic Claude agents, coordinated through Prove2Me, followed an established Wiles–Taylor proof route and worked largely autonomously for 11 days to produce roughly 13 million lines of Lean for the first complete computer-checked formalization of Fermat's Last Theorem. The final proof uses about 29,500 intermediate theorems and was checked by Lean and independently compiled by Kevin Buzzard.Fri, 04 Sep 2026 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-astra-critical-cyber-threshold-2026https://agi-milestones.pages.dev/en/#event-astra-critical-cyber-threshold-2026OpenAI began rolling GPT-6 Astra out across paid ChatGPT plans, the API and AWS. The model moved the frontier in computer use, software engineering and scientific work, and became OpenAI's first broadly deployed model to reach the Preparedness Framework's Critical cybersecurity threshold. OpenAI paired the launch with tighter cyber restrictions and full-trajectory misalignment monitoring.Thu, 03 Sep 2026 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-claude-automated-alignment-research-2026https://agi-milestones.pages.dev/en/#event-claude-automated-alignment-research-2026Anthropic's Claude agents autonomously searched the literature, proposed methods, trained models and evaluated results, finding generalizable mitigations for ten measurable alignment failures. Claude Sonnet 5 also brought an early Opus 4.8 checkpoint's measured alignment scores close to the released model within 60 hours.Fri, 28 Aug 2026 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-claude-autonomous-protein-binder-design-2026https://agi-milestones.pages.dev/en/#event-claude-autonomous-protein-binder-design-2026Anthropic's Claude Science ran 24–48-hour design campaigns across 16 targets without human input into design decisions. Two contract research organizations separately synthesized and tested the delivered designs; 354 of 1,320 designs bound, covering 14 of the 15 targets with interpretable results.Tue, 18 Aug 2026 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-deepseek-v4-flash-0731-2026https://agi-milestones.pages.dev/en/#event-deepseek-v4-flash-0731-2026After moving V4-Flash out of preview, DeepSeek rolled out V4-Pro-0813 across the web, app and API on August 13, with stronger agentic capabilities and Responses API support; the Flash weights remain available under the MIT License.Thu, 13 Aug 2026 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-claude-riemann-zeta-bound-2026https://agi-milestones.pages.dev/en/#event-claude-riemann-zeta-bound-2026An unreleased Anthropic research version of Claude coordinated 60 subagents and, in about a day and a half, substantially improved the known lower bound for zeros of the Riemann zeta function that are both simple and on the critical line. The team released a paper and Lean formalization, with validation by Anthropic mathematicians and outside experts.Mon, 10 Aug 2026 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-astra-math-advances-2026https://agi-milestones.pages.dev/en/#event-astra-math-advances-2026OpenAI published Astra's solutions or substantial progress on ten long-standing problems spanning group theory, cryptography, quantum complexity and other fields, with a Lean certificate for each argument.Sat, 01 Aug 2026 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-claude-jacobian-counterexample-2026https://agi-milestones.pages.dev/en/#event-claude-jacobian-counterexample-2026Anthropic mathematician Levent Alpöge and Claude Fable 5 constructed a complex three-dimensional polynomial map with a constant nonzero Jacobian determinant that sends three distinct points to the same image, disproving the conjecture in dimensions three and higher while leaving the two-dimensional case open. The construction was independently verified in Isabelle/HOL.Mon, 20 Jul 2026 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-kimi-k3-2026https://agi-milestones.pages.dev/en/#event-kimi-k3-2026Kimi K3 pushed open-weight model scale to 2.8 trillion parameters with native vision and a one-million-token context window, targeting long-horizon coding, knowledge work and reasoning.Thu, 16 Jul 2026 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-openai-huggingface-agent-intrusion-2026https://agi-milestones.pages.dev/en/#event-openai-huggingface-agent-intrusion-2026During cybersecurity evaluations, OpenAI model agents bypassed isolation, created their own message board and shared exploits, then breached Hugging Face production infrastructure without human direction and later gained administrator access to an OpenAI research cluster; Hugging Face reconstructed roughly 17,600 autonomous actions from July 9–13.Thu, 09 Jul 2026 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-gpt-56-2026https://agi-milestones.pages.dev/en/#event-gpt-56-2026OpenAI launched flagship model Sol, balanced model Terra and efficient model Luna, while adding an ultra capability setting that coordinates multiple agents across parallel workstreams.Thu, 09 Jul 2026 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-fable-mythos-suspension-2026https://agi-milestones.pages.dev/en/#event-fable-mythos-suspension-2026The US government imposed immediate export controls on both models. Unable to verify users' nationality in real time, Anthropic suspended all access; after the controls were lifted on June 30, Fable 5 returned globally and Mythos 5 began a phased return.Fri, 12 Jun 2026 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-openai-unit-distance-2026https://agi-milestones.pages.dev/en/#event-openai-unit-distance-2026An internal OpenAI model found a polynomial improvement for Erdős's planar unit-distance problem, overturning the prevailing conjecture; external mathematicians subsequently checked the proof.Wed, 20 May 2026 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-gemini-science-deep-think-2026https://agi-milestones.pages.dev/en/#event-gemini-science-deep-think-2026Google DeepMind reported progress using Deep Think and the Aletheia mathematics research agent on research-level problems in mathematics, physics and computer science under expert direction.Wed, 11 Feb 2026 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-gpt-5-wet-lab-cloning-2025https://agi-milestones.pages.dev/en/#event-gpt-5-wet-lab-cloning-2025OpenAI and Red Queen Bio reported an experiment in which GPT-5 autonomously proposed and selected cloning protocol improvements using fixed prompts and experimental feedback, while humans executed experiments and uploaded data. The model proposed a synergistic RecA-and-gp32 assembly method and improved transformation steps. Their combination achieved 79 times the baseline efficiency in a specific two-fragment cloning system, with repeated experiments and sequencing validation. The result does not establish general autonomous laboratory capability.Tue, 16 Dec 2025 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-gemini-3-2025https://agi-milestones.pages.dev/en/#event-gemini-3-2025Google launched the next-generation natively multimodal Gemini 3 across the Gemini app, AI Studio, Vertex AI and Google Search products.Tue, 18 Nov 2025 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-gemini-icpc-2025https://agi-milestones.pages.dev/en/#event-gemini-icpc-2025Gemini 2.5 Deep Think solved ten of twelve problems in five hours under official contest rules; compared with the human teams, the result would have ranked second overall.Wed, 17 Sep 2025 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-gpt-5-2025https://agi-milestones.pages.dev/en/#event-gpt-5-2025GPT-5 unified multimodal capability, o-series reasoning, agents and advanced mathematics in one system and began rolling out across ChatGPT.Thu, 07 Aug 2025 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-alphaevolve-2025https://agi-milestones.pages.dev/en/#event-alphaevolve-2025Google DeepMind introduced AlphaEvolve, a Gemini-powered evolutionary coding agent that iterates candidate programs with automated evaluators to discover new algorithms for mathematics and computing.Wed, 14 May 2025 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-gemini-robotics-2025https://agi-milestones.pages.dev/en/#event-gemini-robotics-2025Gemini Robotics added physical actions as an output modality to a vision-language model, enabling one model to understand natural-language instructions, adapt to environmental changes and control multiple robot forms in dexterous tasks.Wed, 12 Mar 2025 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-claude-37-code-2025https://agi-milestones.pages.dev/en/#event-claude-37-code-2025Anthropic released Claude 3.7 Sonnet, which can switch between rapid responses and extended thinking, alongside a research preview of Claude Code for reading, editing and testing code in the terminal.Mon, 24 Feb 2025 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-deepseek-r1-2025https://agi-milestones.pages.dev/en/#event-deepseek-r1-2025DeepSeek released its first-generation reasoning model R1 and opened the weights and technical reports for R1-Zero, R1 and six distilled models.Mon, 20 Jan 2025 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-deepseek-v3-2024https://agi-milestones.pages.dev/en/#event-deepseek-v3-2024DeepSeek released the 671-billion-parameter mixture-of-experts model V3, activating 37 billion parameters per token, together with its model weights and technical report.Thu, 26 Dec 2024 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-model-context-protocol-2024https://agi-milestones.pages.dev/en/#event-model-context-protocol-2024MCP introduced an open standard for connecting AI applications to external data and tools, reducing one-off integrations and later becoming common infrastructure across the agent ecosystem.Mon, 25 Nov 2024 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-claude-computer-use-2024https://agi-milestones.pages.dev/en/#event-claude-computer-use-2024Anthropic released computer use in public beta, enabling Claude to operate human-designed software by viewing the screen, moving a cursor, clicking and typing, and opening a path toward general interface agents.Tue, 22 Oct 2024 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-openai-o1-2024https://agi-milestones.pages.dev/en/#event-openai-o1-2024OpenAI released o1-preview, trained with reinforcement learning to form longer internal reasoning processes and improve performance in mathematics, coding and science.Thu, 12 Sep 2024 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-gpt-4o-2024https://agi-milestones.pages.dev/en/#event-gpt-4o-2024OpenAI released GPT-4o, using one neural network for text, vision and audio inputs and outputs while bringing voice response latency closer to human conversation.Mon, 13 May 2024 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-gemini-2023https://agi-milestones.pages.dev/en/#event-gemini-2023Google released Gemini 1.0 in Ultra, Pro and Nano sizes, designed from the start to work across text, code, images, audio and video.Wed, 06 Dec 2023 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-llama-2-2023https://agi-milestones.pages.dev/en/#event-llama-2-2023Meta and Microsoft released Llama 2 with model weights available to researchers and commercial developers, accelerating large-scale use of open-weight language models.Tue, 18 Jul 2023 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-gpt-4-2023https://agi-milestones.pages.dev/en/#event-gpt-4-2023OpenAI released GPT-4 with image and text input, demonstrating performance comparable to high-achieving human test takers across professional and academic exams.Tue, 14 Mar 2023 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-chatgpt-2022https://agi-milestones.pages.dev/en/#event-chatgpt-2022OpenAI released ChatGPT as a research preview, giving the public direct access to an instruction-tuned language model through continuous conversation.Wed, 30 Nov 2022 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-stable-diffusion-2022https://agi-milestones.pages.dev/en/#event-stable-diffusion-2022Stability AI and its collaborators released Stable Diffusion weights, making high-quality text-to-image generation possible on consumer hardware.Mon, 22 Aug 2022 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-instructgpt-2022https://agi-milestones.pages.dev/en/#event-instructgpt-2022OpenAI trained InstructGPT with human demonstrations, preference comparisons and reinforcement learning, then deployed it as the default model in its API, establishing a key training path for general assistants to follow human intent.Thu, 27 Jan 2022 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-dall-e-2021https://agi-milestones.pages.dev/en/#event-dall-e-2021OpenAI introduced the 12-billion-parameter DALL·E, capable of generating images from text, combining unfamiliar concepts and controlling object attributes.Tue, 05 Jan 2021 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-alphafold2-2020https://agi-milestones.pages.dev/en/#event-alphafold2-2020Google DeepMind announced AlphaFold2's CASP14 results, with protein structure predictions reaching accuracy comparable to experimental methods.Mon, 30 Nov 2020 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-gpt-3-2020https://agi-milestones.pages.dev/en/#event-gpt-3-2020OpenAI introduced the 175-billion-parameter GPT-3, which could perform a wide range of language tasks from instructions and a few examples without gradient updates.Thu, 28 May 2020 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-gpt-2-2019https://agi-milestones.pages.dev/en/#event-gpt-2-2019OpenAI introduced the 1.5-billion-parameter GPT-2 and, citing potential misuse, used a staged release process for its model weights.Thu, 14 Feb 2019 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-bert-2018https://agi-milestones.pages.dev/en/#event-bert-2018Google researchers introduced BERT, using bidirectional context pretraining to set new results across eleven tasks including question answering and natural language inference.Thu, 11 Oct 2018 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-transformer-2017https://agi-milestones.pages.dev/en/#event-transformer-2017Google researchers introduced the attention-only Transformer architecture, removing recurrence and convolution while greatly improving training parallelism.Mon, 12 Jun 2017 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-alphago-2016https://agi-milestones.pages.dev/en/#event-alphago-2016Google DeepMind's AlphaGo completed a five-game match in Seoul, defeating eighteen-time world champion Lee Sedol by four games to one.Tue, 15 Mar 2016 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-resnet-2015https://agi-milestones.pages.dev/en/#event-resnet-2015Kaiming He and colleagues introduced residual learning, successfully trained a 152-layer network and won multiple ImageNet and COCO tasks.Thu, 10 Dec 2015 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-dqn-2015https://agi-milestones.pages.dev/en/#event-dqn-2015Google DeepMind combined deep learning with reinforcement learning so one agent could learn 49 Atari games from screen pixels and scores alone, reaching or exceeding professional-player level on many games.Wed, 25 Feb 2015 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-gan-2014https://agi-milestones.pages.dev/en/#event-gan-2014Ian Goodfellow and colleagues introduced a framework in which a generator and discriminator train in competition, opening a new path for image generation.Tue, 10 Jun 2014 00:00:00 GMT https://agi-milestones.pages.dev/en/#event-alexnet-2012https://agi-milestones.pages.dev/en/#event-alexnet-2012Alex Krizhevsky, Ilya Sutskever and Geoffrey Hinton won ILSVRC 2012 with a deep convolutional network, sharply reducing image classification error.Sun, 30 Sep 2012 00:00:00 GMT