NVIDIA Nemotron 3.5 Bolt 74B-A9B 24GB is now on Ollama. This vision-language model (VLM) is designed to run everyday agentic tasks locally, using native 2-bit quantization to fit in 24GB of memory and run alongside other applications. ollama run nemotron-3.5-bolt
Why it matters: The stated memory footprint provides a concrete deployment consideration for running local agentic workloads alongside other applications on the same machine.
Meta’s Muse Spark Open is an open-weight mixture-of-experts model with approximately 816 billion parameters, around 42 billion active per token, and a context window of up to one million tokens. Trained for agentic coding, it accepts text, images, audio, and video and generates text. Hugging Face provides inference and deployment recipes, reports its own quantization tests, and demonstrates attention-layer LoRA fine-tuning with TRL while keeping the experts frozen.
Why it matters: The deployment recipes clarify how sparse expert activation reduces per-token computation while retaining substantial memory requirements, helping developers assess hardware needs across checkpoint formats and serving frameworks.
Nvidia released Nemotron 3.5 Super VL, an open multimodal model with 120 billion parameters and 12 billion active per token, for continuously running enterprise agents. Nvidia says it approaches Nemotron 3 Ultra’s accuracy on specified agent tasks with about one-fifth the parameters, and is releasing weights, datasets, and training recipes. The accompanying Nemotron 3.5 Bolt 74B-A9B 24GB targets local agents, while Bolt E4B 4GB is planned for later this month.
Why it matters: The deployment details clarify how quantization, caching, and hardware requirements shape the practical tradeoffs of running continuous enterprise agents on company infrastructure.
NVIDIA released Nemotron 3.5 Super VL, an open 120-billion-parameter model for multimodal agents, and Nemotron 3.5 Bolt 74B-A9B 24GB for local agentic workloads; Bolt E4B 4GB is coming Oct. 23. Super VL targets tool use, business workflows and visual document understanding, with open weights, datasets and training recipes. Bolt 74B-A9B uses native 2-bit quantization to fit in 24GB of memory, while Bolt E4B targets physical and conversational agents.
Why it matters: Open weights, datasets and training recipes give organizations a path to adapting agents to their own workflows while meeting local deployment and data residency requirements.
Project Apertus released Apertus 1.5 in 8B and 70B sizes, adding image and audio understanding, optional thinking, a 262,144-token context window, and improved instruction-following and tool use. The weights were released on Hugging Face on July 24, 2026 under Apache 2.0; the models generate only text, and spoken-language processing remains experimental. The team also publishes training data, recipes, intermediate checkpoints, and reproducible evaluation records.
Why it matters: Published intermediate checkpoints, training mixtures, and evaluation settings provide a concrete basis for investigating how individual training stages change model behavior and for reproducing reported results.
Skywork AI released Matrix-Game 4.0, an open-source game world model that reportedly runs at 480p and 60 FPS on one RTX 5090, with input response under 0.15 seconds and peak VRAM usage of approximately 18.5GiB. A game engine maintains state and executes rules, while a video model generates visuals from structural frames and dynamic captions. Model weights and code are available through https://huggingface.co/Skywork/Matrix-Game-4.0 and https://github.com/SkyworkAI/Matrix-Game.
Why it matters: The engine and video model division offers a transferable design approach for keeping game rules deterministic while using generative models to render interactive worlds.
Ai2 introduced HiRO-ACEv2, an updated climate emulator that generates global 3 km rain, wind, and pressure data at about 1/300th the cost of X-SHiELD. Given coarsened X-SHiELD data, its HiROv2 downscaling model matches high-resolution statistics to within 3%, even for events occurring once in a million data points. The system currently works only with ACE tuned to match X-SHiELD and cannot yet generate high-resolution data for other climates; downloadable models, tutorial notebooks, and a web demo are available.
Why it matters: The comparison with X-SHiELD gives researchers a concrete cost and accuracy reference for extreme weather simulations, with applicability currently limited to the climate represented in training.
Gemini 4 Argon taught a robot to play cornhole just by watching a video. It mapped the physics, coded its own simulation, and taught @Apptronik's Apollo humanoid to sink 7 throws in a row.
Why it matters: The demonstration connects video observation, physics modeling, and a model-written simulation in a concrete example of teaching a humanoid robot a physical task.
Baoyu describes a demonstration in which Gemini 4 Argon used Antigravity to build a Japanese temple complex in Blender from one prompt containing a detailed workflow. The prompt referenced a scene-building skill document and requested architectural plans, four Nano Banana reference images, and coordinated agents that checked renders against those references and iteratively revised the scene. The demonstrator requested animation and programmatically generated bell sounds and ambient audio the next day, making those later additions to the initial modeling workflow.
Why it matters: The workflow offers a transferable approach to complex 3D generation by combining construction rules, visual references, agent coordination, and repeated comparison of rendered results.
Google DeepMind has launched Gemini 4 Argon globally through its API and selected Google products, with access depending on subscription or service tier. Introductory pricing is $2 per million input tokens and $10 per million output tokens, with cached input tokens at 95% off and a 1M unified input-output token limit through the Gemini API. Google reports improvements across coding, professional workflows, scientific research, and robotics, alongside stronger agent safeguards and prompt injection resistance.
Why it matters: Google’s reported gains in code submissions and reductions in rollbacks connect Argon’s engineering capabilities to delivery speed and code stability in internal workflows.
Gemini 4 Argon is rolling out today to Google Ultra and Pro subscribers in @GeminiApp and Search, and to Ultra subscribers in @antigravity. Demis Hassabis praises its knowledge work and coding capabilities, sharing a wizard chess game they asked it to create as an example of its physics understanding and 3D modeling capabilities that reminds him of Battlechess on the Amiga. More information and full availability: https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-argon/
Why it matters: The rollout details distinguish access across GeminiApp, Search, and antigravity, giving subscribers concrete context for where they can try Argon under their current plans.
Meta released Muse Wonder and Muse Wonder Full on October 12, according to Baoyu’s account of the announcement. Developers can access Muse Wonder through Meta Model API or Muse Code, while Full is available to Muse Code subscribers; Meta’s blog and Zuckerberg’s X post differ on consumer availability. The models support a 1 million Token context window and multimodal input with text output, while Meta’s reported benchmarks show strengths in sustained agentic and coding tasks alongside weaker results elsewhere.
Why it matters: The comparison distinguishes stronger results on sustained agentic and coding tasks from broader intelligence rankings, while identifying Meta’s internal benchmarks and limitations in how competing models were tested.
Claude Dashboards and Claude Motion are in beta today. Ask Claude to turn your data into live dashboards and your ideas into animated explainers. https://t.co/V3fU4lMnzp
Why it matters: The announcement presents two ways to use Claude for visual communication, pairing live dashboards built from data with animated explainers built from ideas.
OpenAI says GPT-6 and Intelligent UI are now rolling out in ChatGPT for everyone, offering fast, interactive answers and tools for users’ tasks. Mark Chen encourages users to try Intelligent UI, saying every ChatGPT completion now has the potential to create a custom UI and calling it phenomenal for learning new things. https://twitter.com/OpenAI/status/2107894997538525580
Why it matters: The described custom interfaces could make learning more visual and interactive by placing explanations and task tools directly within the flow of a ChatGPT answer.
OpenAI introduced Intelligent UI in ChatGPT, allowing answers to include interactive visual content such as charts, forms, and clickable buttons. The update is rolling out to all users alongside GPT-6.
Why it matters: The source gives concrete examples of Intelligent UI interactions and its staged rollout, helping readers assess how ChatGPT answers are changing.
Liquid AI released two open-weight models in its d1 decision-model family: d1-3B and the experimental d1-omni-600M.
Why it matters: Liquid AI open-sourced two decision models and reported on-device latency and results across seven benchmarks, providing evidence on the use of small models for structured decisions.
Nano Banana announced Nano Banana 2.1, claiming improvements over its previous models across the board, including visual design, mask-based editing, subject consistency, and more natural-looking imagery. It is available in @GeminiApp, @GoogleAIStudio, and more. Josh Woodward commented, “Mask-based editing is my favorite from today, much more coming soon!” https://twitter.com/NanoBanana/status/2107521202466046453
Why it matters: The announcement frames the upgrade around visual design, editing, subject consistency, and natural appearance, providing a concrete comparison with previous models.
Google DeepMind released EmbeddingGemma 2, a 740M-parameter open model based on the Gemma 4 architecture and licensed under Apache 2.0. It maps text, images, audio, and video into a shared embedding space.
Why it matters: EmbeddingGemma 2 maps text, images, audio, and video into one embedding space and reports concrete on-device memory and vector truncation figures, helping assess local multimodal retrieval.
Mistral released a public preview of Mistral Large 4, a natively multimodal model with 1 trillion parameters and 49 billion active parameters. Its weights will be opened at the end of the month.
Why it matters: Mistral outlined the open-weight release schedule for its 1-trillion-parameter multimodal model and its investment in European compute, allowing comparison with mainstream open models.
Google DeepMind released frontier model Gemini 4 Argon, initially making it available through the Fairwind Program to trusted cybersecurity defenders before gradually expanding access. Pricing starts at $2 per million input tokens and $10 per million output tokens.
Why it matters: Gemini 4 Argon raises the maximum output to 1M tokens and reports multiple benchmark results and internal deployment data, offering a view of its limits on long-running workflows.
