Nemotron 3 Ultra
Nemotron 3 Nano is a compact, open large language model in NVIDIA’s Nemotron 3 family, designed for efficient agentic reasoning, conversational AI, and coding tasks. It uses a hybrid Mixture-of-Experts Mamba-Transformer architecture that activates only a small subset of parameters per token, enabling low-latency inference while maintaining strong accuracy and reasoning performance. It has approximately 31.6 billion total parameters with around 3.2 billion active (3.6 billion including embeddings), allowing it to achieve higher accuracy than previous Nemotron 2 Nano while using less computation per forward pass. Nemotron 3 Nano supports long-context processing of up to one million tokens, enabling it to handle large documents, multi-step workflows, and extended reasoning chains in a single pass. It is designed for high-throughput, real-time execution, excelling in multi-turn conversations, tool calling, and agent-based workflows where tasks require planning, reasoning, and more.
Learn more
Holo2
H Company’s Holo2 model family delivers cost-efficient, high-performance vision-language models tailored for computer-use agents that navigate, localize UI elements, and act across web, desktop, and mobile environments. The series, available in 4 B, 8 B, and 30 B-A3B sizes, builds on their earlier Holo1 and Holo1.5 models, retaining strong UI grounding while significantly enhancing navigation capabilities. Holo2 models use a mixture-of-experts (MoE) architecture, activating only necessary parameters, to optimize efficiency. Trained on curated localization and agent datasets, they can be deployed as drop-in replacements for their predecessors. They support seamless inference in frameworks compatible with Qwen3-VL models and can be integrated into agentic pipelines like Surfer 2. In benchmark testing, Holo2-30B-A3B achieved 66.1% accuracy on ScreenSpot-Pro and 76.1% on OSWorld-G, leading the UI localization category.
Learn more
Sarvam 105B
Sarvam-105B is the flagship large language model in Sarvam’s open source model family, designed to deliver high-performance reasoning, multilingual understanding, and agent-based execution within a single scalable system. Built as a Mixture-of-Experts (MoE) model with approximately 105 billion total parameters, of which only a fraction are activated per token, it achieves strong computational efficiency while maintaining high capability across complex tasks. The model is optimized for advanced reasoning, coding, mathematics, and agentic workflows, making it suitable for tasks that require multi-step problem solving and structured outputs rather than simple conversational responses. Sarvam-105B supports long-context processing of up to around 128K tokens, enabling it to handle large documents, extended conversations, and deep analytical queries without losing coherence.
Learn more
Holo3.1
Holo3.1 is H Company’s family of fast and local computer-use agents, built to operate across web, desktop, and mobile environments while integrating more smoothly into different agent frameworks and deployment targets. Based on the Qwen family, Holo3.1 improves robustness across the environments where computer-use agents are actually deployed, addressing the distribution shifts that appear across mobile devices, alternative agent harnesses, and different execution frameworks. The release expands Holo3’s capabilities beyond browser and desktop control, with major gains in mobile automation, including AndroidWorld improvements from 67% to 79.3% for the 35B-A3B model and from 58% to 71% for the smaller 4B and 9B variants. Holo3.1 also introduces native support for function-calling protocols in addition to structured JSON outputs, helping teams deploy the model inside third-party agent stacks with near-parity between function-calling and native execution.
Learn more