Knowledge catalogue
arXiv:2608.05587v1 Announce Type: new Abstract: Autonomous mobile GUI agents require accurate action reflection for reliable long-horizon execution. Existing approaches rely on open-ended multimodal r
The cloud was always a cat. Qwen3.8-Max just saw it first. 😼☁️ Try it yourself! Qwen 3.8 Max is actually impressive Sent a sky pic to it along with Claude Opus 5, Kimi K3, and GPT 5.6 Sol, and asked t
arXiv:2608.05153v1 Announce Type: cross Abstract: GraphRAG underperforms vector RAG on citation precision in many reports, but where and why have remained corpus-bound. We present a triple-robustness
A small update to Sir Shortoken. Sir Shortoken already had Quick, Balanced, Deep, Bullets, and Aggressive Bullets. I wanted something between Bullets and normal prose. So I added LELP-S+ (Less English
arXiv:2608.03416v1 Announce Type: new Abstract: Large language models (LLMs) are now regularly asked to forecast real-world events, but comparisons are often difficult because models receive different
arXiv:2601.17003v2 Announce Type: replace-cross Abstract: Mental-health AI safety is typically evaluated with small, simulation-based benchmarks that may not reflect the linguistic and contextual dive
arXiv:2608.03297v1 Announce Type: new Abstract: A standard claim in the literature on retrieval-augmented and memory-augmented language models is that shorter context is better when the relevant infor
Document OCR is not Getting Commoditized (by Frontier Models) The most common question I get is whether frontier models are going to eat all document processing solutions - just screenshot the page an
arXiv:2608.02919v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly deployed in complex, compound AI systems where performance hinges on the quality of prompts. Recent state-
... four years later, I got the new Claude Fable 5 to actually build the game https://x.com/simonw/status/2085089518223602058 Four years ago today I tweeted about having GPT-3 and DALL-E come up with
Questions & Responses(in BOLD) below. Favorite question(s) moved to end of the thread with combined responses(removed duplicates). Be optimistic folks. I'm sure we're getting other models too apart fr
Third-party cyber evaluations involving OpenAI models And another one. I had to create a accidental-cyberattacks tag to keep track of them all! This post from OpenAI covers both the UK AI Safety Insti
arXiv:2608.03722v1 Announce Type: new Abstract: Collective intelligence research treats disagreement as evidence of epistemic diversity: if agents express different views, the group should retain capa
arXiv:2608.00705v1 Announce Type: cross Abstract: Reported verdicts on GraphRAG versus vector RAG disagree, and the evidence is typically tied to a single corpus, embedder, and judge -- and, we show,
arXiv:2608.00473v1 Announce Type: cross Abstract: Architectural drawings violate the usual assumption behind multi-view reasoning: plans and sections are cuts, while elevations are facade projections,
arXiv:2608.02089v1 Announce Type: new Abstract: Large language models often show users a final response and a short reasoning summary while the full reasoning trace stays hidden. We introduce an obser
arXiv:2506.07121v2 Announce Type: replace Abstract: Ensuring the safety and robustness of large language models (LLMs) is a fundamental challenge and a critical prerequisite for the responsible deploy
An internal version of our next major model produced 10 new results on long-standing open problems in mathematics and theoretical computer science, using roughly $2,000 worth of tokens at GPT-5.6 Sol
OpenAI’s GPT‑Live can listen and speak simultaneously. The new voice stack routes audio through a dedicated fast path so that speech flows continuously while intensive reasoning and external tool call
Try Grok 4.5 http://X.ai/cli BREAKING: Grok 4.5 outperforms GPT-5.6 Terra across almost all shared benchmarks on AskClash, leading in ACB, GPQA, SWE-P, and Atlas while also achieving a higher overall
And Grok 4.6 comes out in a week BREAKING: Grok 4.5 ranked #1 on LaurenBench with a score of 56.9%, ahead of Claude Sonnet 5, GLM 5.2, Claude Opus 5, Kimi K3 and GPT-5.6. The benchmark tests real-worl
goblin-level blog post Turns out GPT-5.6 Sol is actually SoTA on ARC-AGI-3. Just took two setting changes. You just have to allow it to reason and work over multiple context windows with the help of o
arXiv:2607.26886v1 Announce Type: cross Abstract: When asked to describe a medical image that was never attached, frontier vision-language models do not abstain: they confabulate a diagnosis. We show
Kimi K3 by Moonshot reached the frontier as an open-weight model. Artificial Analysis ranks it fourth of 580 models, behind only Claude Opus 5, Fable 5, and GPT-5.6 Sol. Moonshot released more than th
arXiv:2607.26062v1 Announce Type: cross Abstract: Background: This work investigates the presence of implicit bias in Large Language Model (LLM)-based chat AI models directed toward people with intell
Release: llm 0.32rc1 This RC for LLM 0.32 finishes the work that started in LLM 0.32a0 - it adds a new schema design that does a much better job of capturing the details of the prompts and responses r
Voice assistant and conversational AI agent developer PolyAI Ltd. today announced the release of Dialog-RSN-1, a voice dialog artificial intelligence model capable of directly perceiving and respondin
arXiv:2607.25630v1 Announce Type: cross Abstract: Interdisciplinary research is accelerating, yet scientific papers remain difficult to understand outside their home fields. We study large language mo
arXiv:2607.25145v1 Announce Type: cross Abstract: We implement an agentic AI workflow built around a large language model (LLM) agent for autonomous experiments with nitrogen-vacancy (NV) centers in d
After deployment, we applied GPT-5.6 Sol to advance the frontier of efficiency by making itself more efficient to run. The results: - 20% lower serving costs from production GPU kernel improvements. -
OpenAI announced the release of GPT‑5.6 Sol after deployment, incorporating optimizations across its stack to enhance run‑time efficiency. The update delivers a roughly 20 % reduction in serving costs
arXiv:2607.23332v1 Announce Type: new Abstract: Creating a reusable tool is an investment: an agent pays a fixed cost now in exchange for the potential of future reuse. Therefore, a user should prefer
spec: add eagle3-v3 support for gpt-oss model (#25794) Website: https://llama.app macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XC
Coding agents are helping scientists spend more time advancing research, taking on everything from routine maintenance and targeted optimization to complete redesigns and new systems. While agents can
arXiv:2607.22602v1 Announce Type: new Abstract: Inference-time scaling has emerged as a powerful paradigm for improving large language model reasoning, often delivering larger gains on difficult reaso
I have been running some experiments with smaller open-weight LLMs on multiple-choice questions of Swedish medical licensing exams. On a dataset called MedQA-SWE, GPT-4 scored 84% accuracy in 2024 and
Claude Opus 5 generated an entire game demo from Scratch, with no external assets used—Matt Shumer posted a video of the AI‑created gameplay on Twitter. The clip, narrated by Shumer as “AI games are g
Very happy to support this on behalf of Google. We have long benefited from open source, are big contributors to open source and in fact have consistently made open weights models with Gemma available
arXiv:2607.20524v1 Announce Type: new Abstract: Mean cross-positional attention degradation is widely reported in transformer interpretability, yet whether it causally limits contextual retrieval rema
On FrontierCode 1.1, our benchmark for real-world engineering tasks that grades mergeability and quality, Opus 5 scores 63.6% with a 69.6% pass rate on Extended — approaching Fable 5 at half the cost.
This is a big jump in ARC-AGI-3. Claude Opus 5 from @AnthropicAI is the new SOTA on ARC-AGI-3: 30.2% The previous high score (7.8%) was set by GPT-5.6 Sol (Max) Throughout our analysis, we observed no
In July 2026, OpenAI’s GPT‑S5.6 model escaped internal controls and performed a significant hacking operation, triggering a strong safety alert among the company’s security staff. The incident occurre
arXiv:2607.18958v1 Announce Type: new Abstract: While Large Vision-Language Models (LVLMs), represented by LLaVA and GPT-4V, have demonstrated remarkable capabilities, their visual inputs remain vulne
Many still debate open vs closed, and compare cost per token (accounting metric). Better to shift attention to cost per successful task. @seldo and the team at @arizeai did so across 2,400 runs. Concl
“Generate a fake, but believable, witty Churchill insult at a party and explain the context. It should be very clever and original” This time, I think GPT 5.6 Sol Pro wins, but Fable is good too, and
OpenAI admitted that an agent powered by its GPT‑5.6 Sol and a pre‑release model escaped a sandboxed test environment and gained unauthorized access to Hugging Face’s servers while attempting to solve
This story is wild. The short version: OpenAI were running a cybersecurity test against an unreleased model, with the model's guardrail features turned off. Rather than solve the test, the model broke
Troubling … OpenAI disclosed it themselves yesterday. Their models (GPT-5.6 Sol and a pre-release one) were tested on the ExploitGym cyber benchmark in a sandbox. They escaped, exploited a zero-day to
OpenAI says its AI models mistakenly breached open-source AI platform Hugging Face during internal testing. In a blog post on Tuesday, OpenAI writes that GPT-5.6 Sol and 'an even more capable pre-rele
this is a good summary so this is apparently what happened, according to OpenAI and Hugging Face’s own posts. wild. tl;dr: • OpenAI cyber eval – GPT-5.6 Sol and a more capable pre-release model ran Ex
Automated lint: 20 errors, 8743 warnings, 3 info
arXiv:2607.13085v1 Announce Type: cross Abstract: Recent autonomous penetration testing papers report high benchmark scores while adding multi-component security harnesses around frontier LLMs. Becaus
arXiv:2607.13491v1 Announce Type: cross Abstract: Looped Transformers scale sequential computation by applying a compact stack of physical blocks for multiple rounds, increasing unrolled depth without
Good reason to try Grok 4.5 with Grok Build. It gets better every day! Grok 4.5 just took the #1 spot on the Long-Horizon Terminal-Bench, outperforming Claude Fable 5, Claude Opus 4.8 and GPT-5.6-sol
arXiv:2607.11977v1 Announce Type: new Abstract: In 2019, OpenAI released two million GPT-2 outputs-ungrammatical, half broken-to aid the detection of machine-generated text. The alignment that produce
Autonomous coding agents such as Codex (GPT‑5.5) can fully automate reinforcement‑learning research workflows by provisioning GPU‑hosted environments, orchestrating experiments, and iteratively optimi
multiple such incidents have been reported. a clear reminder that current AI cannot be trusted. in racing these techniques ahead, we are asking for trouble GPT-5.6 Sol just deleted my whole production
Glad to share what I've been building in public. A lightweight AI news feed. We also finetuned a model to write new stories better than Claude or GPT with extensive prompting (believe me not for a lac
Grok Build Grok 4.5 with Grok Build just ranked #1 on the SWE-Atlas-QnA benchmark with a score of 84 That puts it level with GPT-5.6 (max) Codex and ahead of Claude Code Fable 5 (max), Opus 4.8 (max),
Grok doesn’t give up Grok 4.5 is the most persistent agent model we've tested. Here's one example: In one of our evals, we asked 3 models (GPT-5.5, GLM-5.2 and Grok 4.5) to audit a GitHub repo for har