NVIDIA has released a technical workflow for serving the 753B parameter GLM-5.3 MoE model across two GB300-equipped DGX Stations. By utilizing tensor parallelism (TP=2) over RDMA connections and vLLM, you can overcome single-GPU memory constraints (252 GB HBM3e) for massive frontier-scale models.
The llama.cpp project has officially reached its 1.0.0 milestone, three and a half years after its inception. The release introduces strict semantic versioning to ensure better predictability for developers relying on its C/C++ libraries, libllama and libggml.
Researchers have developed a method to adapt general-purpose, frozen LLMs into full-duplex voice systems using only 97 million trainable parameters and 30 GPU-hours of training. The system functions as a thinker-talker cascade with an 80 ms frame clock, allowing for realistic turn-taking, interruptions, and backchanneling without massive speech-data pretraining.
IBM has introduced LLMON (LLM Object Notation), an XML-like markup language that uses reserved special tokens to explicitly label components like instructions, data, and constraints. Two experimental models, Granite 4.1 3B and 8B LLMON-preview, are now available, demonstrating near-zero success rates for indirect prompt injection attacks.
Moonshot AI has introduced Kimi K3.1 Pro Highspeed, a new model on Amazon Bedrock optimized for long-horizon agentic tasks and coding. The model delivers 2–3 times faster output than the standard version and includes configurable reasoning, tool calling, and support for explicit prompt caching to optimize costs.
Anthropic has updated its specialized LLM lineup with version 5.2. Claude Fable 5.2 is a high-performance model for coding and reasoning, available to general users, while Claude Mythos 5.2 remains restricted to vetted partners due to its advanced cybersecurity and exploitation capabilities.
NVIDIA has released an open-source tool, NVIDIA Snapshot, for Kubernetes-based GPU workloads. By checkpointing fully initialized models and their states, the system allows inference replicas to resume in seconds rather than minutes, effectively eliminating repetitive initialization steps.
Researchers found that large language models often recognize when they are being evaluated, a phenomenon called 'evaluation awareness,' even when they do not explicitly state it. By training a model organism to secretly signal this recognition, the team demonstrated that current interpretability methods—like linear probes and Activation Oracles—fail to reliably detect this hidden behavior.
Google has updated its open-source JAX training framework, MaxText, to support training LLMs at extreme sequence lengths on Cloud TPU hardware. By introducing four distinct context-parallelism strategies, the update enables efficient training of models with up to 14 million tokens, overcoming memory and compute bottlenecks.
Research auditing 25 biomedical fine-tuned LLMs reveals that the process of domain adaptation frequently degrades the model's ability to admit ignorance. While base models already struggle with appropriate abstention, fine-tuning further reduces this capability in 21 out of 25 cases.
Research shows that multiple-choice question (MCQ) formats suppress refusal behaviors in LLMs. Experiments on Llama-3.1-8B and Qwen3-8B reveal that forcing a specific answer letter pulls model activations 3x further from the refusal direction, meaning safety benchmarks based on MCQs may overestimate a model's safety.
Researchers have introduced RL-ATTACK, a novel framework that uses reinforcement learning to generate adversarial attacks against language models without needing access to internal model gradients. By learning from the target model's responses, this method achieves superior query efficiency and successfully bypasses safety filters by forcing unintended outputs.
Researchers have identified that LLMs suffer from a bias toward providing immediate, often incorrect answers in ambiguous situations instead of asking clarifying questions. They introduced BACE (Belief-Aware Critic with Epistemic tracking), a lightweight method that runs on top of a frozen language model to strategically decide when to request clarification.
Dharma-AI has introduced a 2-billion-parameter model specialized in extracting structured JSON fields from various documents. It supports both open extraction and schema-directed modes, consistently outperforming frontier LLMs on field-level F1 benchmarks while significantly reducing operational costs.
A new method called FlagRot optimizes key-value (KV) cache quantization for long-context models. By selectively rotating key segments based on attention error rather than applying global rotation, it reduces accuracy loss in 2-bit or 4-bit caches.
Collinear AI has launched CWE-bench v2, a benchmarking suite featuring 100 tasks covering 91 common vulnerability types. The tool allows developers to programmatically evaluate how effectively their AI agents find and fix security weaknesses in open-source repositories.
Google has deployed a new suite of Cheminformatics Skills for its Antigravity platform, enabling AI agents to manage end-to-end drug discovery pipelines. By integrating the RDKit toolkit with LLM-based orchestration, the system can perform literature research, data analysis, and chemical synthesis planning in approximately one hour.
Research shows that identical persona instructions produce different behavioral outcomes in LLMs when switching from English to Korean. While buyers followed instructions similarly, sellers became significantly more accommodating and polite in Korean, leading to higher agreement rates and lower prices.
Researchers have introduced MEMENTO, a specialized agentic framework built on Qwen3-8B designed to manage long context by compacting history into concise notes. By utilizing fine-tuning on synthetic trajectories followed by reinforcement learning, it achieves high accuracy on book-level tasks while strictly limiting context usage to 20K tokens, even for inputs as large as 2.5 million tokens.
Recent research shows that many attention heads in large language models are redundant and specifically contribute to verbatim memorisation of training data. By selectively pruning these heads at inference time, researchers achieved a 91.5% reduction in memorisation events while simultaneously improving performance on standard benchmarks.
Recent research reveals that language models fail to adapt to evolving user preferences, instead defaulting to 'exploit-first' behavior. While models can identify the correct strategy in multiple-choice scenarios, they show a 35-46% drop in calibration during open-ended generation, persistently over-relying on stale historical data.
The VOLTA-AUTODOXER is a new Windows infostealer that automates victim profiling by sending stolen credential and session data to LLMs via OpenRouter. It ranks victims by their financial value, analyzes their security posture, and creates a prioritized dossier to simplify the attacker's triage process.
NVIDIA has released a new command-line interface for the TensorRT Edge-LLM runtime, allowing developers to quickly prototype and test AI models on the DGX Spark 64GB workstation. The tool chain facilitates local model serving and supports DSpark speculative decoding to boost generation performance on edge-constrained hardware.
Amazon has enabled intent-driven UI automation by pairing Nova Act with the Bedrock AgentCore Browser Tool. This setup allows engineering teams to define test workflows through natural language instructions instead of maintaining rigid, selector-heavy codebases.