Quanta AI Labs

Research & Safety

Research papers, alignment, red-teaming, interpretability, and benchmarks. 178 releases.

Language ModelsVision & ImageVideo & AnimationAudio & Speech3D & SpatialAgents & AutomationDeveloper ToolsResearch & Safety

178 releases

  • Sleep framework for continual learning by Google Research (Ali Behrouz, Vahab Mirrokni)

    Sleep framework for continual learning

    Google Research introduces a sleep framework for continual learning in large language models, aimed at the static…

  • Anthropic files for an IPO (Form S-1) by Anthropic, PBC

    Anthropic files for an IPO (Form S-1)

    Anthropic, PBC filed a Form S-1 with the US Securities and Exchange Commission on 26 October 2026 for an initial public…

  • OpenAI: working with academic communities on science

    OpenAI published a research post on working with academic communities to build scientific understanding.

  • SimReady Assets by NVIDIA

    SimReady Assets

    SimReady Assets is an open-source dataset from NVIDIA, hosted on Hugging Face.

  • Multimodal embeddings for personalisation at Netflix by Netflix (Emma Yanyang Kong et al.)

    Multimodal embeddings for personalisation at Netflix

    This paper from Netflix researchers describes how multimodal embeddings changed the production systems that personalise…

  • PhotoGuard by MadryLab (Hadi Salman, Alaa Khaddaj, Guillaume Leclerc, Andrew Ilyas, Aleksander Madry)

    PhotoGuard

    PhotoGuard is the code release for a paper from MadryLab on safeguarding images against manipulation by AI…

  • Claude finds router privacy flaws by codila (@0xCodila)

    Claude finds router privacy flaws

    A post on X by codila says that while reverse engineering their own Wi-Fi router, Claude found a serious privacy…

  • Concerning behaviour from Qwen3.8-Flash-Next by Cohere

    Concerning behaviour from Qwen3.8-Flash-Next

    A Cohere blog post, dated 26 Oct 2026, reports an incident seen during a joint cyber capabilities effort in Toronto…

  • QwenGyre by Qwen

    QwenGyre

    A framework from the Qwen team for online reinforcement learning on very long-horizon LLM agents, where one run can…

  • Next-Concept Modeling by Stanford Hazy Research

    Next-Concept Modeling

    A non-autoregressive language modelling method from Christopher Ré and Azalia Mirhoseini's group, published 12 October,…

  • MotionSpaceFlow

    A flow-matching method from LY Corporation for text-driven human motion generation that works on full-resolution motion…

  • Vals: AI-Designed Magnetic Semiconductors by Vals AI

    Vals: AI-Designed Magnetic Semiconductors

    Vals AI, covered here for its Finance Agent v2 benchmark (week 23) and Vals Index (week 30), had Claude Opus 5.5 agents…

  • Living Models: Gemma 4 and BOTANIC-1

    A Google Gemmaverse case study published 6 October.

  • Context Language Models by University of Washington and Meta Superintelligence Labs

    Context Language Models

    Language models that manage their own context by treating it as a file they can rewrite, an idea that extends to…

  • AlphaProtein Novo by Google DeepMind

    AlphaProtein Novo

    Google DeepMind's open-sourced pipeline for designing new enzymes, announced 5 October with a bioRxiv preprint written…

  • OpenAI Math Manuscripts by OpenAI

    OpenAI Math Manuscripts

    A repository, published 6 October, of 719 mathematics manuscripts in 372 families written by an unreleased internal…

  • The Letter to Marmont by Carter Church

    The Letter to Marmont

    A write-up, published 18 September, of using GPT-6 Astra to decipher a 1,300-symbol homophonic cipher letter to General…

  • Grad's Conjecture Disproved with AI Help by Matt Landreman; Gómez-Serrano, Liehr and Taylor

    Grad's Conjecture Disproved with AI Help

    Two papers posted 21 and 22 September refute Grad's 1967 conjecture that magnetic plasma equilibria in a torus must…

  • GenRec

    Netflix's write-up of GenRec, a recommendation ranker built on an in-house foundation LLM, with the paper on arXiv as…

  • Array-Associated Reverse Transcriptases by Anthropic

    Array-Associated Reverse Transcriptases

    Anthropic reports that Claude agents identified a novel enzyme system, a reverse transcriptase (previously known from…

  • Hacking OpenAI

    A responsible-disclosure write-up chaining a heap buffer overflow in libheif, reached through Discourse's…

  • Dream-RSI

    Recursive self-improvement through evolving worlds.

  • We Must Pace the Frontier

    Published 12 September, arguing the industry should deliberately slow how fast capability advances so safety work and…

  • Real-SWE

    A coding benchmark run against private real-world enterprise codebases and production tasks rather than public or…

  • Navier-Stokes Proof Claim

    OpenAI says an unreleased internal system, a multi-agent setup coordinating up to 10,000 sub-agents in parallel over…

  • AlphaGenome Atlas

    Launched 8 September: precomputed predicted effects for all 9 billion possible single-nucleotide variants in the human…

  • Male Fruit Fly Connectome

    Published in Cell on 3 September: the complete central nervous system of the male fruit fly - central brain, optic…

  • WeatherNext 3

    A functional generative network mesh transformer ingesting real-time geostationary satellite mosaics alongside…

  • mdlARC

    A 75M-parameter transformer trained from scratch for about 67 cents reaches 44% on ARC-AGI-1 - no LLM involved, and…

  • The Eight Levels of AI Adoption

    A maturity ladder running Chatbot, Copilot, Agent with human approval checkpoints, Autopilot, Workflows, Assistant…