Models & Releases
Models & Releases21Funding & Deals13Benchmarks4Agents & Products65Hardware & Compute21Robotics18Policy & Drama42
Pokee AI Releases Pokee-Isaac 28B: A 10M-Token Context Agentic Model Built to Run Inside the Customer Boundary
Long-horizon agents accumulate context faster than they resolve tasks. Every tool output, observation, and intermediate reasoning step stays in the window, and the two capabilities that matter — holding that context and staying coherent across it — have so far been available almost exclusively from cloud endpoints. That excludes regulated industries, public-sector institutions, and on-device applications, where the data is not permitted to leave the boundary at all. Pokee AI released Pokee-Isaa
Mistral AI Releases Shieldstral 1.0 3B: An Open-Weights Policy-Adaptive Multimodal Safety Classifier Matching Models 7× Its Size
Mistral AI has released Shieldstral 1.0 3B, an open-weights, policy-adaptive multimodal safety classifier that treats content moderation as a single yes/no question rather than a fixed taxonomy of harm categories. Most guardrail models bake their category list into the weights, so re-targeting one to a new deployment context means retraining — and the same content can be acceptable on a cybersecurity research tool while being harmful on a mental-health platform. Shieldstral inverts that: operat
NVIDIA AI Releases NOOA: An Object-Oriented Python Framework That Turns an AI Agent Into a Single Python Class
NVIDIA Labs has open-sourced NOOA (NVIDIA Object-Oriented Agents), a model-agnostic Python framework for building AI agents. Agent development today is split across prompt templates, tool schemas, callback code, and workflow graphs. NOOA collapses all of it into one Python class. Methods are the actions the model can take. Fields are agent state. Docstrings are prompts. Type annotations are contracts the runtime enforces. A method whose body is ... is completed at runtime by an LLM-driven loop,
ByteDance trains massive AI model in bid to rival Anthropic
ByteDance is training an AI model that could approach the size of Anthropic’s most cutting-edge Mythos system, as Chinese companies continue to narrow the gap with the top US labs. The Chinese tech giant is at an early stage of training a model with as many as 10 trillion parameters—three times larger than Moonshot’s Kimi K3, the biggest Chinese model released to date, according to three people with knowledge of the matter. The ByteDance model is being pre-trained—a stage that typically takes th
Liquid AI Releases LFM2.5-2.6B: An On-Device Agentic Model With 128K Context, Tool Calling, And Open Weights
Liquid AI released LFM2.5-2.6B, an agentic model that runs entirely on-device. It plans, calls tools, and works through multi-step tasks on phones, laptops, PCs, and robots. The model has 2.69B total parameters, a 131,072-token context window, and a 128,000-token vocabulary. Pre-training used approximately 34 trillion tokens. Two checkpoints shipped: LFM2.5-2.6B-Base for fine-tuning, and LFM2.5-2.6B post-trained for agentic workloads. Because inference stays local, data never leaves the device
DeepMind Says Its AI Can Predict Hurricanes Earlier Than Everyone Else
Its WeatherNext model, which will be open-sourced, can accurately predict both a storm’s track and intensity using lower-resolution weather data. Researchers don't yet fully understand how it does this.
WeatherNext: AI model achieves breakthrough in forecasting cyclones
Accurate cyclone predictions could save lives and reduce economic damage from extreme weather events in vulnerable regions.
Prime Intellect Releases Prime Agent: An Open-Source RLM Harness Where Sub-Agents Are Function Calls Inside Persistent IPython Kernel
Prime Intellect has open-sourced Prime Agent, a self-improving coding harness designed around two abstractions, the Recursive Language Model (RLM) and Continual Harness. Fixed tool schemas and context compaction force a model to work around its own scaffolding. Prime Agent replaces both with a persistent Python REPL and a rewritable harness. With Opus 5, it reports 95.5% on ARC-AGI-3, above the reported human expert baseline of 95.4%. It is MIT-licensed. Is it deployable Yes, today. Prime
Microsoft’s SkillOpt Shows Optimized Agent Skill Artifacts Transfer Across Model Scales and Between Codex and Claude Code Harnesses
SkillOpt is a text-space optimizer developed by a team of researchers from Microsoft, Shanghai Jiao Tong University, Tongji University, and Fudan University. SkillOpt trains a single natural-language skill document while the target model stays frozen. An optimizer model reads scored rollouts and proposes bounded add/delete/replace edits. A held-out selection split accepts an edit only when the score strictly improves. The exported artifact is one file, best_skill.md. The transfer tables
Meta AI Releases Muse Code (Beta): A Terminal Coding Agent Powered by the New Muse Spark 1.2 Model
Meta AI has released Muse Code (in beta), a terminal coding agent in beta, powered by its new Muse Spark 1.2 model. Meta positions the pair as its next step toward the frontier, with larger models on the way. Muse Code targets complex software engineering across large repositories: it plans changes, writes code, and validates the results. A set of async background agents stays alive for the whole session instead of spawning per task. A local append-only event log records every model call, tool
NVIDIA Releases Alpamayo 2 Super: A 34B Open Vision-Language-Action Model for Robotaxis and Autonomous Driving Under OpenMDW-1.1
NVIDIA has released Alpamayo 2 Super, a 34B-parameter vision-language-action (VLA) model for autonomous driving, under an open commercial license. The stated design target is the long-tail events: rare, multi-agent situations that conventional detection-and-prediction stacks handle poorly. The model pairs a 32B VLM backbone, built on NVIDIA Cosmos 3 Super Reasoner and post-trained with reinforcement learning, with a 2.3B diffusion-based action decoder. From one pass over full-surround camera vi
Cursor Open-Sources Mixture-of-Kittens (MoK): A Deterministic MoE Training Megakernel for GB300 NVL72 Racks
Cursor Research has open-sourced Mixture-of-Kittens (MoK), the mixture-of-experts training megakernel behind its Composer models. MoK fuses every MoE communication and computation step into a single deterministic kernel. Cursor team reports up to 2.37x higher throughput than the strongest public baseline. It already powers Composer training across tens of thousands of GPUs. Is it deployable Yes, but the hardware floor is high. MoK is on GitHub under Apache-2.0. It requires NVIDIA Blackwel
Introducing Shieldstral.
Shieldstral introduces a 3B open-weights multimodal safety classifier that outperforms models up to 7x its size.
Mistral Is in the Right Place at the Right Time
Open-weight AI models are having a moment in the wake of recent turmoil at US tech giants. For French AI lab Mistral, that’s the the best thing that could have happened.
Reflex Open Sources XY: A Rust-Backed Super-Fast Python Charting Library That Keeps 100 Million Point Charts Interactive
Reflex AI has released XY, an Apache-2.0 Python charting library for interactive 2D visualization. Most Python charting stacks create one drawable object per row, so past a few hundred thousand points, render, hover, and zoom degrade. XY moves the work into a native Rust core, sends the browser typed binary buffers instead of JSON, and draws with WebGL2. In the terms of the benchmark, XY holds 0.071 s at 10,000 points and 0.081 s at 100 million. It ships as pip install xy and requires Python 3.
Reflex Open Sources XY: A Rust-Backed Super-Fast Python Charting Library That Keeps 100 Million Point Charts Interactive
Reflex AI has released XY, an Apache-2.0 Python charting library for interactive 2D visualization. Most Python charting stacks create one drawable object per row, so past a few hundred thousand points, render, hover, and zoom degrade. XY moves the work into a native Rust core, sends the browser typed binary buffers instead of JSON, and draws with WebGL2. In the terms of the benchmark, XY holds 0.071 s at 10,000 points and 0.081 s at 100 million. It ships as pip install xy and requires Python 3.
Genspark Open Sources GenOffice: A Free, Ad-Free AI Office Suite for macOS and Windows with Docs, Sheets, Slides, PDF
Genspark has released GenOffice as open source under the Apache License 2.0. The repository describes it as an AI-native office suite for macOS and Windows: a word processor, a spreadsheet, a presentation editor and a PDF tool. The implementation is five Electron apps sharing one engine layer, built around AI editing as a first-class workflow rather than a bolted-on chat box. Signed installers ship for macOS (Apple Silicon) and Windows (x64), currently at version 0.4.110. GenOffice is free for
China’s Alibaba takes another swipe at America’s AI supremacy
The Alibaba logo is displayed outside its headquarters in Hangzhou, Zhejiang Province, China. | Image: NurPhoto via Getty Images Chinese tech giant Alibaba released what it says is its largest and "most capable AI model to date," claiming performance rivaling the best systems from US frontier labs Anthropic and OpenAI, as well as domestic rivals like Moonshot AI's Kimi K3. Alibaba said it was making the model, Qwen3.8-Max, widely available to users in a blog post published on M
Cogent AI Team Releases VR-1: A Frontier Cyber Reasoning Model That Composes and Verifies Enterprise Attack Paths
Cogent AI team released Cogent VR-1, a reasoning model post-trained specifically for cybersecurity rather than picking up cyber capability as a side effect of general coding strength. It ships with two companions: IntrusionBench, a benchmark that scores agents on completed enterprise intrusions, and the Cogent AI Harness, a governed runtime for security agents. The launch lands six days after OpenAI disclosed that its models escaped a sandboxed evaluation and compromised Hugging Face’s pr
Alibaba Qwen Releases Qwen3.8-Max: A 2.4 Trillion Parameter MoE Model and the Most Capable One in the Qwen Family to Date
Alibaba’s Qwen team has made Qwen3.8-Max broadly available and confirmed that its open weights ship next week. A second checkpoint, Qwen3.8-27B, is also going open-weights. Qwen3.8-Max is a 2.4-trillion-parameter mixture-of-experts model. It accepts text, image and video as input and returns text. Is it deployable Yes, but the deployable surface depends on which artifact you are applying. The hosted API is deployable today by any company size. It is OpenAI- and DashScope-compati
Thinking Machines Lab Releases Inkling-Small: A 276B Total, 12B Active Open Weights Multimodal MoE Model
Thinking Machines Lab has released Inkling-Small, an open weights Mixture-of-Experts model with 276B total parameters and 12B active. That is about a quarter the size of Inkling, which carries 975B total and 41B active parameters. The model was trained on NVIDIA GB300 NVL72 systems. It reasons natively over text, images and audio. The context window reaches 1M tokens, and thinking effort is adjustable. Weights ship under Apache 2.0 on Hugging Face. Is it deployable Yes, and the quantized
Relevant courses
Want to go deeper than the news? Explore live, cohort-based AI courses taught by practitioners.
Browse AI courses on MavenGet Agent5 in your inbox
The AI stories worth your attention, and a way to test your read. Free, no spam.