AI Milestones

First documented case of an AI system independently discovering a critical operating system kernel vulnerability.

OpenAI announces preparation to file for initial public offering, marking a major corporate milestone for the AI industry leader.

Anthropic announces major infrastructure expansion to Colossus2 cluster using NVIDIA's GB200 chips for next-generation model training.

ByteDance releases Lance, an open-source 3B parameter model combining image/video generation and understanding capabilities in a single architecture.

An OpenAI model successfully disproved a central conjecture in discrete geometry, demonstrating advanced mathematical reasoning capabilities.

Google releases Gemini 3.5 Flash, representing a major advancement in their flagship model series with enhanced action capabilities.

Largest comparative study shows AI system surpasses doctors in clinical diagnosis and decision-making using real emergency department data.

First unified multimodal model performing visual understanding and generation directly from pixel embeddings without vision encoders.

First documented case of AI assistance enabling solution to a major unsolved mathematical problem.

UAE becomes first country to commit to running half of government operations via autonomous AI systems.

Google commits up to $40 billion in cash and compute resources to Anthropic in massive AI partnership deal.

DeepSeek releases V4 Pro, a major new flagship model representing significant advancement in their model capabilities.

OpenAI releases GPT-5.5, a major flagship model update with significant capability improvements.

Anthropic receives $5B from Amazon with commitment to $100B in cloud spending, marking major industry partnership.

Apollo foundation model trained on 25 billion medical records from 7.2 million patients across 28 modalities, representing the largest healthcare AI model.

荣耀's 'Lightning' robot completed Beijing half-marathon in 50:26, faster than human world record, marking breakthrough in autonomous athletic performance.

Anthropic CEO meets with White House officials over national security concerns regarding Claude Mythos's autonomous vulnerability discovery capabilities.

New framework enables verification-aware speculative decoding with step-level error detection without external reward models, improving multi-step reasoning accuracy.

Breakthrough in 3D policy learning with transformer-based encoder solving training instabilities and overfitting issues.

Novel reasoning-driven framework treats sign language translation as cross-modal reasoning task with latent thought sequences.

OpenAI introduces multi-turn approach for precise GUI interaction with visual feedback and error correction.

System enabling autonomous AI agents to sustain coherent ML research progress across comprehension, implementation, and experimentation over days.

rDPO framework introduces instance-specific rubrics for fine-grained visual reasoning preference optimization in multimodal tasks.

Monthly-refreshed benchmark testing whether LLMs can find known security vulnerabilities in real repository codebases with sandboxed exploration.

Berkeley researchers demonstrate systematic ways to break top AI agent benchmarks, highlighting fundamental evaluation methodology issues.

Open-source visual web agent with transparent training data and methodology for autonomous web navigation tasks.

Anthropic introduces ClawBench, a comprehensive evaluation framework testing AI agents on 153 everyday online tasks across 144 live platforms.

Research breakthrough addressing agents' meta-cognitive deficits in arbitrating between internal knowledge and external tool usage.

Meta introduces Muse Spark, positioning it as a step toward personal superintelligence capabilities for individual users.

Research breakthrough allows full-precision training of 100+ billion parameter language models on a single GPU, dramatically reducing training costs.

Anthropic releases specialized Claude model variant focused on advanced cybersecurity capabilities with detailed system card documentation.

Google releases Gemma-4 series with any-to-any and image-text-to-text capabilities across multiple parameter sizes (4B-31B).

Claude successfully wrote a complete FreeBSD remote kernel RCE exploit with root shell, demonstrating advanced cybersecurity capabilities.

Original Alibaba Qwen technical lead publishes influential essay on transitioning from reasoning to agentic thinking paradigms.

New benchmark designed to measure artificial general intelligence through novel reasoning tasks, addressing limitations of previous AI evaluation methods.

First multimodal framework combining video, tactile sensing, and action prediction for contact-rich physical interactions.

OpenAI introduces framework to accelerate multimodal agent reasoning through speculative perception and planning.

First AI system confirmed to solve an open mathematical research problem, marking breakthrough in AI mathematical reasoning capabilities.

First demonstration of a 400 billion parameter language model running natively on a mobile device, showcasing dramatic advances in on-device AI.

Researchers discover discrete 3-4 layer 'reasoning circuits' in transformers that can be duplicated to dramatically improve logical deduction performance without training.

Research introduces framework enabling language models to continuously improve from real-world deployment experience rather than offline training only.

Nvidia introduces purpose-built CPU architecture specifically designed for agentic AI workloads, marking hardware specialization for autonomous agents.

Investment bank warns of imminent AI breakthrough driven by rapid computing expansion that could strain power grids and disrupt jobs globally.

Legendary programmer John Carmack publicly disputes OpenAI and other labs' aggressive AGI timelines, stating 'We Are Not on the Brink of AGI' with significant implications for industry investment.

Anthropic's Claude models now support 1 million token context windows in general availability, enabling processing of extremely long documents.

First desktop agent that learns tasks from single demonstrations across GUI apps, browsers, terminals, and messaging tools in unified sessions.

Nvidia announces major strategic shift with $26 billion investment in open-source AI models over five years, competing directly with OpenAI and other closed-source providers.