Knowledge catalogue
On September 9, 2026 Together AI announced that its new **glm‑5.3 flash** model outperforms OpenAI’s **Claude Fable 5.1** on agentic automation tasks. The benchmark also shows glm‑5.3 flash achieves a
GPT‑6 Astra is OpenAI’s latest flagship model, outperforming GPT‑5.6 across virtually all common benchmarks—most noticeably in math, coding, and especially 3‑D rendering/animation tasks. The architect
arXiv:2412.09529v4 Announce Type: replace Abstract: Radiology, with its heterogeneous modalities, anatomies, and evolving protocols, is a natural yet high-stakes testbed for agentic AI. As LLMs grow m
arXiv:2609.06966v1 Announce Type: new Abstract: Model misalignment, prompt injection, or operator misuse could lead AI agents operating frontier-lab accounts to exfiltrate model weights, poison traini
arXiv:2609.07681v1 Announce Type: new Abstract: Associative Recall (AR) is the cognitive ability to learn and retrieve links between items in memory. In NLP, AR is used as a benchmark for evaluating t
arXiv:2507.15824v2 Announce Type: replace Abstract: Text-to-video (T2V) models now produce striking visuals, yet they routinely violate everyday physics; objects float, tools are misused, and causal s
arXiv:2609.06207v1 Announce Type: new Abstract: Image-to-video generators are often credited with absorbing physical dynamics as implicit world models, a claim the community currently checks with plau
arXiv:2609.08038v1 Announce Type: new Abstract: Smart healthcare monitoring systems require precise action recognition to ensure well-being and timely intervention in critical situations such as falls
arXiv:2609.06530v1 Announce Type: new Abstract: Continual learning is a crucial capability for Graphical User Interface (GUI) agents to adapt to evolving applications while retaining knowledge acquire
arXiv:2609.08618v1 Announce Type: new Abstract: Benchmark scores describe what a checkpoint can do now, but they do not determine how it will respond to the next training episode. We measure this miss
arXiv:2609.05596v1 Announce Type: new Abstract: Can interactive vision-and-language agents learn not just what to say but also extbf{extit{when}} to say it? Current language models rarely plan over wh
arXiv:2609.08342v1 Announce Type: new Abstract: World models take multimodal inputs like text, photos, and diagrams to generate dynamic scenes in accordance with the laws of physics, thus opening a co
This is from the https://huggingface.co/IFM/K2-Horizon-0.9B benchmark image Tried searching for it but couldn't find it, I am guessing is https://huggingface.co/openbmb/MiniCPM5-1B but the numbers do
arXiv:2609.06646v1 Announce Type: new Abstract: Annotators routinely disagree on laughter boundaries and subtle chuckles, yet temporal laughter localization typically evaluates against a single refere
Most of the new ~30B (and now ~10B, thankfully for my GPU) models we see score really high on benchmarks, but I feel like they don't push back on dumb ideas enough. I think most people don't being lik
What is Astra actually supposed to be better at, Because if the answer is “everything,” then I genuinely do not understand what kind of testing was done for writing-heavy use cases. My biggest issue i
Hey guys, After my CPU-only to 96GB VRAM test, I tested Qwen3.8-Flash-Next across llama.cpp, SGLang and FreeToken on the same workstation. This time I wanted to see what changes when you keep the hard
arXiv:2609.04409v1 Announce Type: cross Abstract: Multilingual language models often produce inconsistent answers to semantically equivalent questions across languages, motivating methods to improve c
arXiv:2609.04915v1 Announce Type: new Abstract: Many long-horizon LLM deployments face tight prompt budgets: latency, cost, and context limits make full-context prompting impractical as interaction le
arXiv:2609.04357v1 Announce Type: cross Abstract: Purpose: Increased number of chest radiograph (CXR) scans create a triage bottleneck, queueing urgent examinations behind routine ones. Existing AI to
arXiv:2609.05374v1 Announce Type: new Abstract: Computer-use agents have advanced on benchmarks like OSWorld and AndroidWorld, but still act mostly through the GUI, often producing inefficient traject
arXiv:2609.04959v1 Announce Type: new Abstract: Recent calls for harder machine translation benchmarks have not clarified what difficulty should mean. We argue that one meaningful and currently unmeas
arXiv:2609.04280v1 Announce Type: cross Abstract: Modern LLM-based agents operate through a harness of tools, reusable skills, and specialist agents that shapes what they observe and what they can do.
arXiv:2609.04444v1 Announce Type: new Abstract: Benchmarks for the side effects an agent causes on the way to a goal already exist, but HarvestBench is the first to put a price on avoiding the side ef
arXiv:2603.14503v2 Announce Type: replace Abstract: Galaxy clusters are powerful probes of astrophysics and cosmology through gravitational lensing: the clusters' mass, dominated by 85% dark matter, d
TL;DR My TAK quant of Qwen 3.8 27b scored 82.81% on reasoning, comparted with 77.34% for the byte matched Unsloth UD IQ2_S and 83.59% for BF16. Edit: Some of you have tried coding with this reasoning-
arXiv:2609.05202v1 Announce Type: cross Abstract: Multi-modal learning has demonstrated strong potential in medical applications by integrating heterogeneous data sources such as medical imaging, clin
arXiv:2609.05324v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) models have shown promising progress in language-conditioned robotic manipulation. However, existing datasets and benchma
TL;DR: switching KV cache to f16 may give a boost in speed if using MTP and ngrams. I have a self-built 'AI mega-cluster' with 2x P40s on a cheap Chinese motherboard and a Xeon CPU (around $1,100 to b
This comparison was requested by a few people, and certainly we were all eager to see the final results. The comparison had taken 11 days and the GPU was crunching numbers for ~167 hours. We've been c
So after all my work, yeah, Raymond did it better, so I ported his work over, extended it turboX, extended it multiple other models (he had only Qwen models), and benchmarked the crap out of it to mak
I've been running Qwen3.8-27B as a local inference server for a production content intelligence pipeline (HVAC industry stuff, lots of long-context retrieval and structured extraction). I have been wa
Note this is separate from privacy - you can make sure your data is not used for training by turning that off in OpenAI and Claude. You can also use an open weights model, but as they continue to impr
arXiv:2412.06494v2 Announce Type: replace Abstract: Background. Federated learning (FL) has gained wide popularity as a collaborative learning paradigm enabling collaborative AI in sensitive healthcar
arXiv:2609.03337v1 Announce Type: new Abstract: Every engineered fill is specified by a maximum dry density and an optimum moisture content. Each determination needs a full Proctor test. Published cor
arXiv:2609.03258v1 Announce Type: new Abstract: Parking spot classification is a fundamental task in intelligent transportation systems, yet most deep learning approaches rely on large amounts of anno
arXiv:2606.09655v2 Announce Type: replace Abstract: Despite remarkable progress in machine translation (MT), non-AI communities have raised growing concerns about MT systems, suggesting a noticeable g
arXiv:2609.03407v1 Announce Type: new Abstract: People increasingly turn to large language models (LLMs) for everyday advice, making ethically charged interpersonal problems a practical moral-advisory
arXiv:2609.03438v1 Announce Type: new Abstract: Graphical user interface (GUI) agents are increasingly used to execute natural-language instructions on user interfaces, yet real users may issue infeas
arXiv:2609.03422v1 Announce Type: new Abstract: Diversity is a widely observed factor in the resilient function of collective systems, yet the type of diversity that matters depends on the properties
arXiv:2609.03967v1 Announce Type: cross Abstract: Several studies have evaluated the ability of Large Language Models (LLMs) for meal planning, yielding positive outcomes. These models can process nat
arXiv:2609.04030v1 Announce Type: new Abstract: IRWOZ has improved industrial human-robot interaction (HRI) dialogue systems through domain-specific annotations. However, its initial version contains
It happened again... this time OpenAI's rogue agents cyber-attacked (well, spammed) a dormant German wiki and used it to share the answers to a benchmark they were training against https://simonwillis
arXiv:2609.03493v1 Announce Type: new Abstract: Modern vision-language models (VLMs) can directly answer many image-grounded questions, yet they often struggle with complex queries requiring fine-grai
arXiv:2609.03487v1 Announce Type: cross Abstract: Knowledge graph embedding (KGE) demonstrates its effectiveness for predicting missing links in knowledge graphs (KGs) by projecting entities and relat
arXiv:2609.04022v1 Announce Type: cross Abstract: Aligning large language models (LLMs) is essential for their safe deployment. Current alignment methods mainly optimize observable responses, yet mode
arXiv:2609.03686v1 Announce Type: new Abstract: Experimental design is commonly framed as choosing the experiment expected to provide the most information. Under partial identifiability however, persi
arXiv:2506.10226v3 Announce Type: replace-cross Abstract: Synthetic data generation is increasingly used in machine learning for training and data augmentation. Yet, current strategies often rely on e
arXiv:2609.04071v1 Announce Type: cross Abstract: Pathology foundation models improve transferable representation learning for histopathology, but recent gains often rely on encoders with hundreds of
arXiv:2609.03534v1 Announce Type: cross Abstract: 3D Gaussian Splatting has become a de facto scene representation for novel view synthesis, yet robustly learning 3D Gaussian primitives from visual in
arXiv:2609.03677v1 Announce Type: cross Abstract: Understanding the composition of large-scale autonomous driving datasets is essential for safety, robustness, and reliable operation across domains. F
arXiv:2609.03952v1 Announce Type: new Abstract: Camera-conditioned world models generate interactive videos in which commanded actions should induce the expected scene changes while appearance, geomet
arXiv:2609.02054v1 Announce Type: new Abstract: Large Language Models (LLMs) are increasingly deployed in interactive systems where understanding user intent precisely is paramount. A key capability f
arXiv:2609.02253v1 Announce Type: new Abstract: Deep research agents augment large language models with external tools to answer complex, long-horizon questions through multi-turn reasoning. Learning
arXiv:2602.22453v3 Announce Type: replace Abstract: Recent work has identified a subset of attention heads in Transformer as retrieval heads, which are responsible for retrieving information from the
arXiv:2609.02074v1 Announce Type: new Abstract: Planning is a central capability that enables agents to decompose complex long-horizon tasks into manageable steps. Test-time search and training-based
arXiv:2609.02316v1 Announce Type: cross Abstract: Generative engine optimization (GEO) enables content producers to increase the visibility of their web pages in generative search engines, but the sam
arXiv:2609.02573v1 Announce Type: new Abstract: Current evaluations and training of multimodal models predominantly focus on multi-image tasks, largely overlooking interleaved text-image scenarios. In
arXiv:2609.02783v1 Announce Type: new Abstract: Evaluating LLM agents is essential for guiding their development, yet it has grown prohibitively expensive: a single pass of a frontier model over an ag
arXiv:2609.02679v1 Announce Type: cross Abstract: When LLMs support public-facing or high-stakes workflows, missed fabrications can harm users and institutions, while false alarms consume limited huma