Alternatives to Gemma 2

NVIDIA Nemotron 3.5 Bolt is a series of efficient AI models developed by NVIDIA for running autonomous agents and conversational applications directly on local hardware. The family consists of Bolt 74B-A9B 24GB and Bolt E4B 4GB, which address different performance and memory requirements. Bolt 74B-A9B uses a mixture-of-experts architecture with 74 billion total parameters and approximately 9 billion active parameters to balance reasoning capabilities with computational efficiency. Native 2-bit quantization enables the model to operate within a 24GB memory footprint on supported NVIDIA systems. Its agentic capabilities include web research, code generation, software debugging, information retrieval, and multistep task execution. The model can be integrated with local AI agent frameworks and development tools to complete assignments without sending application workloads to external inference services. Bolt E4B is optimized for fast conversational responses, multistep interactions, and embedded AI experiences such as virtual assistants and interactive game characters. NVIDIA supports local deployment of the Bolt family across compatible GeForce RTX, RTX PRO, DGX Spark, and Jetson hardware, depending on the model variant. Bolt 74B-A9B is available with open weights through Hugging Face and supports inference tools including llama.cpp, Ollama, LM Studio, Unsloth, and vLLM.

Muse Spark Open is an open-weight foundation model developed by Meta to support multimodal reasoning, autonomous agents, and complex software development tasks. The model features 816 billion total parameters with approximately 42 billion active parameters and a context window supporting up to one million tokens. Its architecture is designed to process multimodal information while maintaining context across lengthy workflows, extensive codebases, and multistep assignments. Muse Spark Open provides capabilities for software engineering, code generation, debugging, application development, and automated testing. Its agentic functionality supports external tool interactions, task planning, computer-based workflows, and execution of complex objectives over extended sessions. The model can run locally using two Apple M5 Ultra Mac Studio systems with the MLX framework, enabling organizations to process information without relying on hosted inference services. Developers can deploy the model through SGLang, vLLM, TensorRT-LLM, Hugging Face Transformers, llama.cpp, and other supported inference engines. Customization is supported through PyTorch and TorchTitan, allowing organizations to adapt the model to specialized applications and workflows. Muse Spark Open is distributed under the Muse Model License and is available through Hugging Face, major cloud model repositories, and hosted API providers including Together AI, Fireworks AI, and OpenRouter.

The Gemini 1.5 Pro AI model represents a pinnacle in language modeling, engineered to produce remarkably precise, context-sensitive, and human-like replies suitable for a wide range of uses. Its innovative neural framework allows it to excel in tasks involving natural language comprehension, generation, and reasoning. This model has been meticulously fine-tuned for adaptability, making it capable of handling diverse activities such as content creation, coding, data analysis, and intricate problem-solving. Its sophisticated algorithms provide a deep understanding of language, allowing for smooth adjustments to various domains and conversational tones. Prioritizing both scalability and efficiency, the Gemini 1.5 Pro is designed to cater to both small applications and large-scale enterprise deployments, establishing itself as an invaluable asset for driving productivity and fostering innovation. Moreover, its ability to learn from user interactions enhances its performance, making it even more effective in real-world scenarios.

The Gemini 1.5 Flash AI model represents a sophisticated, high-speed language processing system built to achieve remarkable speed and immediate responsiveness. It is specifically crafted for environments that necessitate swift and timely performance, integrating an optimized neural framework with the latest technological advancements to ensure outstanding efficiency while maintaining precision. This model is particularly well-suited for high-velocity data processing needs, facilitating quick decision-making and effective multitasking, making it perfect for applications such as chatbots, customer support frameworks, and interactive platforms. Its compact yet robust architecture allows for efficient deployment across various settings, including cloud infrastructures and edge computing devices, thus empowering organizations to enhance their operational capabilities with unparalleled flexibility. Furthermore, the model’s design prioritizes both performance and scalability, ensuring it meets the evolving demands of modern businesses.

OLMo 2 represents a collection of completely open language models created by the Allen Institute for AI (AI2), aimed at giving researchers and developers clear access to training datasets, open-source code, reproducible training methodologies, and thorough assessments. These models are trained on an impressive volume of up to 5 trillion tokens and compete effectively with top open-weight models like Llama 3.1, particularly in English academic evaluations. A key focus of OLMo 2 is on ensuring training stability, employing strategies to mitigate loss spikes during extended training periods, and applying staged training interventions in the later stages of pretraining to mitigate weaknesses in capabilities. Additionally, the models leverage cutting-edge post-training techniques derived from AI2's Tülu 3, leading to the development of OLMo 2-Instruct models. To facilitate ongoing enhancements throughout the development process, an actionable evaluation framework known as the Open Language Modeling Evaluation System (OLMES) was created, which includes 20 benchmarks that evaluate essential capabilities. This comprehensive approach not only fosters transparency but also encourages continuous improvement in language model performance.

Gemma 3, launched by Google, represents a cutting-edge AI model constructed upon the Gemini 2.0 framework, aimed at delivering superior efficiency and adaptability. This innovative model can operate seamlessly on a single GPU or TPU, which opens up opportunities for a diverse group of developers and researchers. Focusing on enhancing natural language comprehension, generation, and other AI-related functions, Gemma 3 is designed to elevate the capabilities of AI systems. With its scalable and robust features, Gemma 3 aspires to propel the evolution of AI applications in numerous sectors and scenarios, potentially transforming the landscape of technology as we know it.

Mistral AI has launched two cutting-edge models designed for on-device computing and edge applications, referred to as "les Ministraux": Ministral 3B and Ministral 8B. These innovative models redefine the standards of knowledge, commonsense reasoning, function-calling, and efficiency within the sub-10B category. They are versatile enough to be utilized or customized for a wide range of applications, including managing complex workflows and developing specialized task-focused workers. Capable of handling up to 128k context length (with the current version supporting 32k on vLLM), Ministral 8B also incorporates a unique interleaved sliding-window attention mechanism to enhance both speed and memory efficiency during inference. Designed for low-latency and compute-efficient solutions, these models excel in scenarios such as offline translation, smart assistants that don't rely on internet connectivity, local data analysis, and autonomous robotics. Moreover, when paired with larger language models like Mistral Large, les Ministraux can effectively function as streamlined intermediaries, facilitating function-calling within intricate multi-step workflows, thereby expanding their applicability across various domains. This combination not only enhances performance but also broadens the scope of what can be achieved with AI in edge computing.

Olmo 3 represents a comprehensive family of open models featuring variations with 7 billion and 32 billion parameters, offering exceptional capabilities in base performance, reasoning, instruction, and reinforcement learning, while also providing transparency throughout the model development process, which includes access to raw training datasets, intermediate checkpoints, training scripts, extended context support (with a window of 65,536 tokens), and provenance tools. The foundation of these models is built upon the Dolma 3 dataset, which comprises approximately 9 trillion tokens and utilizes a careful blend of web content, scientific papers, programming code, and lengthy documents; this thorough pre-training, mid-training, and long-context approach culminates in base models that undergo post-training enhancements through supervised fine-tuning, preference optimization, and reinforcement learning with accountable rewards, resulting in the creation of the Think and Instruct variants. Notably, the 32 billion Think model has been recognized as the most powerful fully open reasoning model to date, demonstrating performance that closely rivals that of proprietary counterparts in areas such as mathematics, programming, and intricate reasoning tasks, thereby marking a significant advancement in open model development. This innovation underscores the potential for open-source models to compete with traditional, closed systems in various complex applications.

On September 17, 2024, Mistral AI revealed a series of significant updates designed to improve both the accessibility and efficiency of their AI products. Among these updates was the introduction of a complimentary tier on "La Plateforme," their serverless platform that allows for the tuning and deployment of Mistral models as API endpoints, which gives developers a chance to innovate and prototype at zero cost. In addition, Mistral AI announced price reductions across their complete model range, highlighted by a remarkable 50% decrease for Mistral Nemo and an 80% cut for Mistral Small and Codestral, thereby making advanced AI solutions more affordable for a wider audience. The company also launched Mistral Small v24.09, a model with 22 billion parameters that strikes a favorable balance between performance and efficiency, making it ideal for various applications such as translation, summarization, and sentiment analysis. Moreover, they released Pixtral 12B, a vision-capable model equipped with image understanding features, for free on "Le Chat," allowing users to analyze and caption images while maintaining strong text-based performance. This suite of updates reflects Mistral AI's commitment to democratizing access to powerful AI technologies for developers everywhere.

Introducing Mistral NeMo, our latest and most advanced small model yet, featuring a cutting-edge 12 billion parameters and an expansive context length of 128,000 tokens, all released under the Apache 2.0 license. Developed in partnership with NVIDIA, Mistral NeMo excels in reasoning, world knowledge, and coding proficiency within its category. Its architecture adheres to industry standards, making it user-friendly and a seamless alternative for systems currently utilizing Mistral 7B. To facilitate widespread adoption among researchers and businesses, we have made available both pre-trained base and instruction-tuned checkpoints under the same Apache license. Notably, Mistral NeMo incorporates quantization awareness, allowing for FP8 inference without compromising performance. The model is also tailored for diverse global applications, adept in function calling and boasting a substantial context window. When compared to Mistral 7B, Mistral NeMo significantly outperforms in understanding and executing detailed instructions, showcasing enhanced reasoning skills and the ability to manage complex multi-turn conversations. Moreover, its design positions it as a strong contender for multi-lingual tasks, ensuring versatility across various use cases.

In honor of Archimedes, whose 2311th anniversary we celebrate this year, we are excited to introduce our inaugural Mathstral model, a specialized 7B architecture tailored for mathematical reasoning and scientific exploration. This model features a 32k context window and is released under the Apache 2.0 license. Our intention behind contributing Mathstral to the scientific community is to enhance the pursuit of solving advanced mathematical challenges that necessitate intricate, multi-step logical reasoning. The launch of Mathstral is part of our wider initiative to support academic endeavors, developed in conjunction with Project Numina. Much like Isaac Newton during his era, Mathstral builds upon the foundation laid by Mistral 7B, focusing on STEM disciplines. It demonstrates top-tier reasoning capabilities within its category, achieving remarkable results on various industry-standard benchmarks. Notably, it scores 56.6% on the MATH benchmark and 63.47% on the MMLU benchmark, showcasing the performance differences by subject between Mathstral 7B and its predecessor, Mistral 7B, further emphasizing the advancements made in mathematical modeling. This initiative aims to foster innovation and collaboration within the mathematical community.

Phi-4 is an advanced small language model (SLM) comprising 14 billion parameters, showcasing exceptional capabilities in intricate reasoning tasks, particularly in mathematics, alongside typical language processing functions. As the newest addition to the Phi family of small language models, Phi-4 illustrates the potential advancements we can achieve while exploring the limits of SLM technology. It is currently accessible on Azure AI Foundry under a Microsoft Research License Agreement (MSRLA) and is set to be released on Hugging Face in the near future. Due to significant improvements in processes such as the employment of high-quality synthetic datasets and the careful curation of organic data, Phi-4 surpasses both comparable and larger models in mathematical reasoning tasks. This model not only emphasizes the ongoing evolution of language models but also highlights the delicate balance between model size and output quality. As we continue to innovate, Phi-4 stands as a testament to our commitment to pushing the boundaries of what's achievable within the realm of small language models.

Gemma 4 is an advanced AI model developed by Google as part of its Gemini architecture, designed to deliver strong performance while remaining accessible to developers. The model is optimized to run on a single GPU or TPU, allowing more organizations and researchers to experiment with powerful AI technology. Gemma 4 improves natural language understanding and generation, making it suitable for applications such as chatbots, text analysis, and automated content creation. Its architecture enables the model to process complex language patterns while maintaining efficient computational performance. Developers can integrate Gemma 4 into various AI projects that require intelligent text processing or conversational capabilities. The model is designed with scalability in mind, allowing it to support both research experiments and production systems. By offering high-performance AI in a more accessible format, Gemma 4 lowers the barrier for developing sophisticated AI solutions. Its flexibility makes it useful for industries ranging from technology and education to business automation. Researchers can also use the model to explore new AI techniques and improve language processing systems. Overall, Gemma 4 represents a step forward in making powerful AI models easier to deploy and use.

Gemma represents a collection of cutting-edge, lightweight open models that are built upon the same research and technology underlying the Gemini models. Created by Google DeepMind alongside various teams at Google, the inspiration for Gemma comes from the Latin word "gemma," which translates to "precious stone." In addition to providing our model weights, we are also offering tools aimed at promoting developer creativity, encouraging collaboration, and ensuring the ethical application of Gemma models. Sharing key technical and infrastructural elements with Gemini, which stands as our most advanced AI model currently accessible, Gemma 2B and 7B excel in performance within their weight categories when compared to other open models. Furthermore, these models can conveniently operate on a developer's laptop or desktop, demonstrating their versatility. Impressively, Gemma not only outperforms significantly larger models on crucial benchmarks but also maintains our strict criteria for delivering safe and responsible outputs, making it a valuable asset for developers.

Introducing Gemma 3n, our cutting-edge open multimodal model designed specifically for optimal on-device performance and efficiency. With a focus on responsive and low-footprint local inference, Gemma 3n paves the way for a new generation of intelligent applications that can be utilized on the move. It has the capability to analyze and respond to a blend of images and text, with plans to incorporate video and audio functionalities in the near future. Developers can create smart, interactive features that prioritize user privacy and function seamlessly without an internet connection. The model boasts a mobile-first architecture, significantly minimizing memory usage. Co-developed by Google's mobile hardware teams alongside industry experts, it maintains a 4B active memory footprint while also offering the flexibility to create submodels for optimizing quality and latency. Notably, Gemma 3n represents our inaugural open model built on this revolutionary shared architecture, enabling developers to start experimenting with this advanced technology today in its early preview. As technology evolves, we anticipate even more innovative applications to emerge from this robust framework.

CodeGemma represents an impressive suite of efficient and versatile models capable of tackling numerous coding challenges, including middle code completion, code generation, natural language processing, mathematical reasoning, and following instructions. It features three distinct model types: a 7B pre-trained version designed for code completion and generation based on existing code snippets, a 7B variant fine-tuned for translating natural language queries into code and adhering to instructions, and an advanced 2B pre-trained model that offers code completion speeds up to twice as fast. Whether you're completing lines, developing functions, or crafting entire segments of code, CodeGemma supports your efforts, whether you're working in a local environment or leveraging Google Cloud capabilities. With training on an extensive dataset comprising 500 billion tokens predominantly in English, sourced from web content, mathematics, and programming languages, CodeGemma not only enhances the syntactical accuracy of generated code but also ensures its semantic relevance, thereby minimizing mistakes and streamlining the debugging process. This powerful tool continues to evolve, making coding more accessible and efficient for developers everywhere.

TranslateGemma is an innovative collection of open machine translation models created by Google, based on the Gemma 3 architecture, which facilitates communication between individuals and systems in 55 languages by providing high-quality AI translations while ensuring efficiency and wide deployment options. Offered in sizes of 4 B, 12 B, and 27 B parameters, TranslateGemma encapsulates sophisticated multilingual functionalities into streamlined models that are capable of functioning on mobile devices, consumer laptops, local systems, or cloud infrastructure, all without compromising on precision or performance; assessments indicate that the 12 B variant can exceed the capabilities of larger baseline models while requiring less computational power. The development of these models involved a distinct two-phase fine-tuning approach that integrates high-quality human and synthetic translation data, using reinforcement learning to enhance translation accuracy across a variety of language families. This innovative methodology ensures that users benefit from an array of languages while experiencing swift and reliable translations.

Mistral Small 3.1 represents a cutting-edge, multimodal, and multilingual AI model that has been released under the Apache 2.0 license. This upgraded version builds on Mistral Small 3, featuring enhanced text capabilities and superior multimodal comprehension, while also accommodating an extended context window of up to 128,000 tokens. It demonstrates superior performance compared to similar models such as Gemma 3 and GPT-4o Mini, achieving impressive inference speeds of 150 tokens per second. Tailored for adaptability, Mistral Small 3.1 shines in a variety of applications, including instruction following, conversational support, image analysis, and function execution, making it ideal for both business and consumer AI needs. The model's streamlined architecture enables it to operate efficiently on hardware such as a single RTX 4090 or a Mac equipped with 32GB of RAM, thus supporting on-device implementations. Users can download it from Hugging Face and access it through Mistral AI's developer playground, while it is also integrated into platforms like Gemini Enterprise Agent Platform, with additional accessibility on NVIDIA NIM and more. This flexibility ensures that developers can leverage its capabilities across diverse environments and applications.

DataGemma signifies a groundbreaking initiative by Google aimed at improving the precision and dependability of large language models when handling statistical information. Released as a collection of open models, DataGemma utilizes Google's Data Commons, a comprehensive source of publicly available statistical information, to root its outputs in actual data. This project introduces two cutting-edge methods: Retrieval Interleaved Generation (RIG) and Retrieval Augmented Generation (RAG). The RIG approach incorporates real-time data verification during the content generation phase to maintain factual integrity, while RAG focuses on acquiring pertinent information ahead of producing responses, thereby minimizing the risk of inaccuracies often referred to as AI hallucinations. Through these strategies, DataGemma aspires to offer users more reliable and factually accurate answers, representing a notable advancement in the effort to combat misinformation in AI-driven content. Ultimately, this initiative not only underscores Google's commitment to responsible AI but also enhances the overall user experience by fostering trust in the information provided.

DiffusionGemma is an innovative open model that investigates text diffusion, representing a remarkably rapid method for generating text. Released under the Apache 2.0 license, this 26 billion parameter Mixture of Experts (MoE) model advances beyond the usual sequential token generation typical of autoregressive models. Instead, it produces entire blocks of text at once, achieving text generation speeds that are up to four times faster on GPUs. Drawing from the parameter efficiency of the Gemma 4 family and Gemini Diffusion research, DiffusionGemma incorporates a unique diffusion head that enhances generation speed significantly. It is particularly aimed at researchers and developers looking to optimize speed-sensitive, interactive local workflows, including in-line editing, swift iterations, and non-linear narrative forms. By reallocating the decode bottleneck from memory bandwidth to computational power, it can produce over 1,000 tokens per second on a single NVIDIA H100 and more than 700 tokens per second on an NVIDIA GeForce RTX 5090. This breakthrough allows for a new level of efficiency in text generation that could reshape various applications in natural language processing.

Mistral 7B is a language model with 7.3 billion parameters that demonstrates superior performance compared to larger models such as Llama 2 13B on a variety of benchmarks. It utilizes innovative techniques like Grouped-Query Attention (GQA) for improved inference speed and Sliding Window Attention (SWA) to manage lengthy sequences efficiently. Released under the Apache 2.0 license, Mistral 7B is readily available for deployment on different platforms, including both local setups and prominent cloud services. Furthermore, a specialized variant known as Mistral 7B Instruct has shown remarkable capabilities in following instructions, outperforming competitors like Llama 2 13B Chat in specific tasks. This versatility makes Mistral 7B an attractive option for developers and researchers alike.

PaliGemma 2 represents the next step forward in tunable vision-language models, enhancing the already capable Gemma 2 models by integrating visual capabilities and simplifying the process of achieving outstanding performance through fine-tuning. This advanced model enables users to see, interpret, and engage with visual data, thereby unlocking an array of innovative applications. It comes in various sizes (3B, 10B, 28B parameters) and resolutions (224px, 448px, 896px), allowing for adaptable performance across different use cases. PaliGemma 2 excels at producing rich and contextually appropriate captions for images, surpassing basic object recognition by articulating actions, emotions, and the broader narrative associated with the imagery. Our research showcases its superior capabilities in recognizing chemical formulas, interpreting music scores, performing spatial reasoning, and generating reports for chest X-rays, as elaborated in the accompanying technical documentation. Transitioning to PaliGemma 2 is straightforward for current users, ensuring a seamless upgrade experience while expanding their operational potential. The model's versatility and depth make it an invaluable tool for both researchers and practitioners in various fields.

Mistral Small 4 is a next-generation open-source AI model created by Mistral AI to deliver powerful reasoning, coding, and multimodal capabilities within a single unified architecture. The model merges features from several specialized systems, including Magistral for advanced reasoning, Pixtral for multimodal processing, and Devstral for agentic software development tasks. It supports both text and image inputs, enabling applications such as conversational AI, document analysis, and visual data interpretation. The model is built using a mixture-of-experts design with 128 experts, allowing efficient scaling while maintaining strong performance across diverse tasks. Users can adjust the model’s reasoning behavior through a configurable parameter that toggles between lightweight responses and deeper analytical processing. Mistral Small 4 also provides a large context window that enables it to handle long conversations, detailed documents, and complex reasoning chains. Compared with earlier versions, the model offers improved performance, reduced latency, and higher throughput for real-time applications. Developers can integrate it with popular machine learning frameworks such as Transformers, vLLM, and llama.cpp. The model’s open-source Apache 2.0 license allows organizations to fine-tune and customize it for specialized use cases. By combining efficiency, flexibility, and multimodal intelligence, Mistral Small 4 provides a versatile foundation for building advanced AI-powered applications.

Introducing Gemma, your innovative AI companion designed to spark creativity and streamline your workflow. With Gemma, you can brainstorm fresh ideas, enhance current designs, and handle repetitive tasks, allowing you to concentrate on what truly inspires you. Whether you need assistance crafting compelling headlines, engaging body text, or memorable brand names, Gemma is here to help. Additionally, Gemma can generate highly realistic images that can be easily resized and modified to suit your needs. Available around the clock, Gemma’s user-friendly interface opens the door to a multitude of AI models and integrates seamlessly with the creative tools you already use. With a focus on learning from your input and preferences, Gemma offers unique suggestions and valuable insights that can elevate your projects. Installing Gemma on your desktop is a breeze, enabling you to access this powerful tool across various files and applications effortlessly. Say goodbye to the intimidating blank page, as Gemma’s cutting-edge algorithms empower your artistic pursuits and transform your visions into reality. You’ll find that collaborating with Gemma is like having a creative partner by your side, ready to explore new horizons together.

The latest iteration of the open-source AI model, which can be fine-tuned and deployed in various environments, is now offered in multiple versions, including 1B, 3B, 11B, and 90B, alongside the option to continue utilizing Llama 3.1. Llama 3.2 comprises a series of large language models (LLMs) that come pretrained and fine-tuned in 1B and 3B configurations for multilingual text only, while the 11B and 90B models accommodate both text and image inputs, producing text outputs. With this new release, you can create highly effective and efficient applications tailored to your needs. For on-device applications, such as summarizing phone discussions or accessing calendar tools, the 1B or 3B models are ideal choices. Meanwhile, the 11B or 90B models excel in image-related tasks, enabling you to transform existing images or extract additional information from images of your environment. Overall, this diverse range of models allows developers to explore innovative use cases across various domains.

Introducing the next iteration of our open-source large language model, this version features model weights along with initial code for the pretrained and fine-tuned Llama language models, which span from 7 billion to 70 billion parameters. The Llama 2 pretrained models have been developed using an impressive 2 trillion tokens and offer double the context length compared to their predecessor, Llama 1. Furthermore, the fine-tuned models have been enhanced through the analysis of over 1 million human annotations. Llama 2 demonstrates superior performance against various other open-source language models across multiple external benchmarks, excelling in areas such as reasoning, coding capabilities, proficiency, and knowledge assessments. For its training, Llama 2 utilized publicly accessible online data sources, while the fine-tuned variant, Llama-2-chat, incorporates publicly available instruction datasets along with the aforementioned extensive human annotations. Our initiative enjoys strong support from a diverse array of global stakeholders who are enthusiastic about our open approach to AI, including companies that have provided valuable early feedback and are eager to collaborate using Llama 2. The excitement surrounding Llama 2 signifies a pivotal shift in how AI can be developed and utilized collectively.

Qwen2.5-VL-32B represents an advanced AI model specifically crafted for multimodal endeavors, showcasing exceptional skills in reasoning related to both text and images. This iteration enhances the previous Qwen2.5-VL series, resulting in responses that are not only of higher quality but also more aligned with human-like formatting. The model demonstrates remarkable proficiency in mathematical reasoning, nuanced image comprehension, and intricate multi-step reasoning challenges, such as those encountered in benchmarks like MathVista and MMMU. Its performance has been validated through comparisons with competing models, often surpassing even the larger Qwen2-VL-72B in specific tasks. Furthermore, with its refined capabilities in image analysis and visual logic deduction, Qwen2.5-VL-32B offers thorough and precise evaluations of visual content, enabling it to generate insightful responses from complex visual stimuli. This model has been meticulously optimized for both textual and visual tasks, making it exceptionally well-suited for scenarios that demand advanced reasoning and understanding across various forms of media, thus expanding its potential applications even further.

EmbeddingGemma 2 is a versatile and lightweight multimodal embedding model that facilitates the mapping of text, code, images, video, and audio into a unified embedding space, which is useful for applications such as search, retrieval, classification, routing, and RAG. It is constructed on the Gemma 4 framework and distributed under the Apache 2.0 license, featuring an impressive 740 million parameters while being fine-tuned for efficient on-device inference. Its flexible architecture allows for the use of only 270 million parameters for text-centric tasks, and it includes additional vision and audio encoders for comprehensive multimodal capabilities. Furthermore, the innovative Matryoshka Representation Learning technique enables developers to compress output vectors from 768 dimensions down to 512, 256, or even 128 dimensions, effectively reducing the storage and memory demands for local vector databases. The model is equipped with an 8K-token context window, providing the capability to handle up to 5.5 minutes of audio, 29 images, 58 video frames, or various combinations of these inputs seamlessly on local hardware. This adaptability makes it particularly valuable for developers seeking to enhance their applications with rich multimedia integration.

GLM-4.5V-Flash is a vision-language model that is open source and specifically crafted to integrate robust multimodal functionalities into a compact and easily deployable framework. It accommodates various types of inputs including images, videos, documents, and graphical user interfaces, facilitating a range of tasks such as understanding scenes, parsing charts and documents, reading screens, and analyzing multiple images. In contrast to its larger counterparts, GLM-4.5V-Flash maintains a smaller footprint while still embodying essential visual language model features such as visual reasoning, video comprehension, handling GUI tasks, and parsing complex documents. This model can be utilized within “GUI agent” workflows, allowing it to interpret screenshots or desktop captures, identify icons or UI components, and assist with both automated desktop and web tasks. While it may not achieve the performance enhancements seen in the largest models, GLM-4.5V-Flash is highly adaptable for practical multimodal applications where efficiency, reduced resource requirements, and extensive modality support are key considerations. Its design ensures that users can harness powerful functionalities without sacrificing speed or accessibility.

Muse Glimmer is an open-weights model featuring 30 billion parameters, developed by Meta Superintelligence Labs, and is fine-tuned for continuous local agent operations. Its compact design allows it to function on a standard Mac or PC equipped with a single consumer GPU, making it ideal for various tasks such as local agent management, function calling, programming, and LLM-as-a-judge evaluations without reliance on cloud services or internet connectivity. This innovative model integrates advanced capabilities such as long-horizon execution, accurate tool invocation, multimodal comprehension, extended memory for context, and effective instruction following. It is proficient in accomplishing end-to-end tasks as an agent, maintains the ability to engage in multi-step reasoning over lengthy processes, can recover gracefully from failed or unanticipated tool engagements, and interprets interleaved text and images using a specialized perception encoder designed for analyzing screenshots, graphs, and document files. Furthermore, Muse Glimmer is compatible with OpenClaw and other orchestration frameworks, allowing for adjustable reasoning efforts, and has been developed with a diverse dataset encompassing over 100 languages. The model's versatility ensures that it can adapt to various applications, thus enhancing its utility in different domains.

Falcon 2 11B is a versatile AI model that is open-source, supports multiple languages, and incorporates multimodal features, particularly excelling in vision-to-language tasks. It outperforms Meta’s Llama 3 8B and matches the capabilities of Google’s Gemma 7B, as validated by the Hugging Face Leaderboard. In the future, the development plan includes adopting a 'Mixture of Experts' strategy aimed at significantly improving the model's functionalities, thereby advancing the frontiers of AI technology even further. This evolution promises to deliver remarkable innovations, solidifying Falcon 2's position in the competitive landscape of artificial intelligence.

MedGemma is an innovative suite of Gemma 3 variants specifically designed to excel in the analysis of medical texts and images. This resource empowers developers to expedite the creation of AI applications focused on healthcare. Currently, MedGemma offers two distinct variants: a multimodal version with 4 billion parameters and a text-only version featuring 27 billion parameters. The 4B version employs a SigLIP image encoder, which has been meticulously pre-trained on a wealth of anonymized medical data, such as chest X-rays, dermatological images, ophthalmological images, and histopathological slides. Complementing this, its language model component is trained on a wide array of medical datasets, including radiological images and various pathology visuals. MedGemma 4B can be accessed in both pre-trained versions, denoted by the suffix -pt, and instruction-tuned versions, marked by the suffix -it. For most applications, the instruction-tuned variant serves as the optimal foundation to build upon, making it particularly valuable for developers. Overall, MedGemma represents a significant advancement in the integration of AI within the medical field.

Phi-4-reasoning is an advanced transformer model featuring 14 billion parameters, specifically tailored for tackling intricate reasoning challenges, including mathematics, programming, algorithm development, and strategic planning. Through a meticulous process of supervised fine-tuning on select "teachable" prompts and reasoning examples created using o3-mini, it excels at generating thorough reasoning sequences that optimize computational resources during inference. By integrating outcome-driven reinforcement learning, Phi-4-reasoning is capable of producing extended reasoning paths. Its performance notably surpasses that of significantly larger open-weight models like DeepSeek-R1-Distill-Llama-70B and nears the capabilities of the comprehensive DeepSeek-R1 model across various reasoning applications. Designed for use in settings with limited computing power or high latency, Phi-4-reasoning is fine-tuned with synthetic data provided by DeepSeek-R1, ensuring it delivers precise and methodical problem-solving. This model's ability to handle complex tasks with efficiency makes it a valuable tool in numerous computational contexts.

LFM2 represents an advanced series of on-device foundation models designed to provide a remarkably swift generative-AI experience across a diverse array of devices. By utilizing a novel hybrid architecture, it achieves decoding and pre-filling speeds that are up to twice as fast as those of similar models, while also enhancing training efficiency by as much as three times compared to its predecessor. These models offer a perfect equilibrium of quality, latency, and memory utilization suitable for embedded system deployment, facilitating real-time, on-device AI functionality in smartphones, laptops, vehicles, wearables, and various other platforms, which results in millisecond inference, device durability, and complete data sovereignty. LFM2 is offered in three configurations featuring 0.35 billion, 0.7 billion, and 1.2 billion parameters, showcasing benchmark results that surpass similarly scaled models in areas including knowledge recall, mathematics, multilingual instruction adherence, and conversational dialogue assessments. With these capabilities, LFM2 not only enhances user experience but also sets a new standard for on-device AI performance.

The Nemotron 3 Nano stands out as the tiniest model within NVIDIA's Nemotron 3 lineup, specifically designed for agentic AI tasks that require robust reasoning and conversational skills while maintaining cost-effective inference. This hybrid Mamba-Transformer Mixture-of-Experts model boasts 3.2 billion active parameters, 3.6 billion when including embeddings, and a total of 31.6 billion parameters. NVIDIA asserts that this model offers greater accuracy compared to its predecessor, the Nemotron 2 Nano, all while utilizing less than half of the parameters during each forward pass, thus enhancing efficiency without compromising on performance. It is also claimed to surpass the accuracy of both GPT-OSS-20B and Qwen3-30B-A3B-Thinking-2507 across various widely-used benchmarks. With an 8K input and 16K output setting utilizing a single H200, the model achieves an inference throughput that is 3.3 times greater than that of Qwen3-30B-A3B and 2.2 times that of GPT-OSS-20B. Additionally, the Nemotron 3 Nano is capable of handling context lengths of up to 1 million tokens, further establishing its superiority over GPT-OSS-20B and Qwen3-30B-A3B-Instruct-2507. This remarkable combination of features positions it as a leading choice for advanced AI applications that demand both precision and efficiency.

Falcon-7B is a causal decoder-only model comprising 7 billion parameters, developed by TII and trained on an extensive dataset of 1,500 billion tokens from RefinedWeb, supplemented with specially selected corpora, and it is licensed under Apache 2.0. What are the advantages of utilizing Falcon-7B? This model surpasses similar open-source alternatives, such as MPT-7B, StableLM, and RedPajama, due to its training on a remarkably large dataset of 1,500 billion tokens from RefinedWeb, which is further enhanced with carefully curated content, as evidenced by its standing on the OpenLLM Leaderboard. Additionally, it boasts an architecture that is finely tuned for efficient inference, incorporating technologies like FlashAttention and multiquery mechanisms. Moreover, the permissive nature of the Apache 2.0 license means users can engage in commercial applications without incurring royalties or facing significant limitations. This combination of performance and flexibility makes Falcon-7B a strong choice for developers seeking advanced modeling capabilities.

Seed2.0 Mini represents the most compact version of ByteDance's Seed2.0 line of versatile multimodal agent models, crafted for efficient high-throughput inference and dense deployment, while still embodying the essential strengths found in its larger counterparts regarding multimodal understanding and instruction adherence. This Mini variant, alongside Pro and Lite siblings, is particularly fine-tuned for handling high-concurrency and batch generation tasks, proving itself ideal for scenarios where the ability to process numerous requests simultaneously is as crucial as its overall capability. In line with other models in the Seed2.0 family, it showcases notable improvements in visual reasoning and motion perception, excels at extracting structured information from intricate inputs such as text and images, and effectively carries out multi-step instructions. However, in exchange for enhanced inference speed and cost efficiency, it sacrifices some degree of raw reasoning power and output quality, ensuring that it remains a practical option for various applications. As a result, Seed2.0 Mini strikes a balance between performance and efficiency, appealing to developers seeking to optimize their systems for scalable solutions.

The o3-mini-high model developed by OpenAI enhances artificial intelligence reasoning capabilities by improving deep problem-solving skills in areas such as programming, mathematics, and intricate tasks. This model incorporates adaptive thinking time and allows users to select from various reasoning modes—low, medium, and high—to tailor performance to the difficulty of the task at hand. Impressively, it surpasses the o1 series by an impressive 200 Elo points on Codeforces, providing exceptional efficiency at a reduced cost while ensuring both speed and precision in its operations. As a notable member of the o3 family, this model not only expands the frontiers of AI problem-solving but also remains user-friendly, offering a complimentary tier alongside increased limits for Plus subscribers, thereby making advanced AI more widely accessible. Its innovative design positions it as a significant tool for users looking to tackle challenging problems with enhanced support and adaptability.

Phi-4-reasoning-plus is an advanced reasoning model with 14 billion parameters, enhancing the capabilities of the original Phi-4-reasoning. It employs reinforcement learning for better inference efficiency, processing 1.5 times the number of tokens compared to its predecessor, which results in improved accuracy. Remarkably, this model performs better than both OpenAI's o1-mini and DeepSeek-R1 across various benchmarks, including challenging tasks in mathematical reasoning and advanced scientific inquiries. Notably, it even outperforms the larger DeepSeek-R1, which boasts 671 billion parameters, on the prestigious AIME 2025 assessment, a qualifier for the USA Math Olympiad. Furthermore, Phi-4-reasoning-plus is accessible on platforms like Azure AI Foundry and HuggingFace, making it easier for developers and researchers to leverage its capabilities. Its innovative design positions it as a top contender in the realm of reasoning models.

EmbeddingGemma is a versatile multilingual text embedding model with 308 million parameters, designed to be lightweight yet effective, allowing it to operate seamlessly on common devices like smartphones, laptops, and tablets. This model, based on the Gemma 3 architecture, is capable of supporting more than 100 languages and can handle up to 2,000 input tokens, utilizing Matryoshka Representation Learning (MRL) for customizable embedding sizes of 768, 512, 256, or 128 dimensions, which balances speed, storage, and accuracy. With its GPU and EdgeTPU-accelerated capabilities, it can generate embeddings in a matter of milliseconds—taking under 15 ms for 256 tokens on EdgeTPU—while its quantization-aware training ensures that memory usage remains below 200 MB without sacrificing quality. Such characteristics make it especially suitable for immediate, on-device applications, including semantic search, retrieval-augmented generation (RAG), classification, clustering, and similarity detection. Whether used for personal file searches, mobile chatbot functionality, or specialized applications, its design prioritizes user privacy and efficiency. Consequently, EmbeddingGemma stands out as an optimal solution for a variety of real-time text processing needs.

We are excited to announce the launch of Phi-2, a language model featuring 2.7 billion parameters that excels in reasoning and language comprehension, achieving top-tier results compared to other base models with fewer than 13 billion parameters. In challenging benchmarks, Phi-2 competes with and often surpasses models that are up to 25 times its size, a feat made possible by advancements in model scaling and meticulous curation of training data. Due to its efficient design, Phi-2 serves as an excellent resource for researchers interested in areas such as mechanistic interpretability, enhancing safety measures, or conducting fine-tuning experiments across a broad spectrum of tasks. To promote further exploration and innovation in language modeling, Phi-2 has been integrated into the Azure AI Studio model catalog, encouraging collaboration and development within the research community. Researchers can leverage this model to unlock new insights and push the boundaries of language technology.

Falcon-40B is a causal decoder-only model consisting of 40 billion parameters, developed by TII and trained on 1 trillion tokens from RefinedWeb, supplemented with carefully selected datasets. It is distributed under the Apache 2.0 license. Why should you consider using Falcon-40B? This model stands out as the leading open-source option available, surpassing competitors like LLaMA, StableLM, RedPajama, and MPT, as evidenced by its ranking on the OpenLLM Leaderboard. Its design is specifically tailored for efficient inference, incorporating features such as FlashAttention and multiquery capabilities. Moreover, it is offered under a flexible Apache 2.0 license, permitting commercial applications without incurring royalties or facing restrictions. It's important to note that this is a raw, pretrained model and is generally recommended to be fine-tuned for optimal performance in most applications. If you need a version that is more adept at handling general instructions in a conversational format, you might want to explore Falcon-40B-Instruct as a potential alternative.

Seed2.0 Lite belongs to the Seed2.0 lineup from ByteDance, which encompasses versatile multimodal AI agent models engineered to tackle intricate, real-world challenges while maintaining a harmonious balance between efficiency and performance. This model boasts superior multimodal comprehension and instruction-following skills compared to its predecessors in the Seed series, allowing it to effectively interpret and analyze text, visual components, and structured data for use in production environments. Positioned as a mid-sized option within the family, Lite is fine-tuned to provide high-quality results with quick responsiveness at a reduced cost and faster inference times than the Pro version, while also enhancing the capabilities of earlier models. Consequently, it is well-suited for applications that demand consistent reasoning, extended context comprehension, and the execution of multimodal tasks without necessitating the utmost raw performance levels. Moreover, this accessibility makes Seed2.0 Lite an attractive choice for developers seeking efficiency alongside capabilities in their AI solutions.

The Mixtral 8x7B model is an advanced sparse mixture of experts (SMoE) system that boasts open weights and is released under the Apache 2.0 license. This model demonstrates superior performance compared to Llama 2 70B across various benchmarks while achieving inference speeds that are six times faster. Recognized as the leading open-weight model with a flexible licensing framework, Mixtral also excels in terms of cost-efficiency and performance. Notably, it competes with and often surpasses GPT-3.5 in numerous established benchmarks, highlighting its significance in the field. Its combination of accessibility, speed, and effectiveness makes it a compelling choice for developers seeking high-performing AI solutions.

It enables efficient training on Amazon Elastic Compute Cloud (Amazon EC2) Trn1 instances powered by AWS Trainium. Additionally, for model deployment, it facilitates both high-performance and low-latency inference utilizing AWS Inferentia-based Amazon EC2 Inf1 instances along with AWS Inferentia2-based Amazon EC2 Inf2 instances. With the Neuron SDK, users can leverage widely-used frameworks like TensorFlow and PyTorch to effectively train and deploy machine learning (ML) models on Amazon EC2 Trn1, Inf1, and Inf2 instances with minimal alterations to their code and no reliance on vendor-specific tools. The integration of the AWS Neuron SDK with these frameworks allows for seamless continuation of existing workflows, requiring only minor code adjustments to get started. For those involved in distributed model training, the Neuron SDK also accommodates libraries such as Megatron-LM and PyTorch Fully Sharded Data Parallel (FSDP), enhancing its versatility and scalability for various ML tasks. By providing robust support for these frameworks and libraries, it significantly streamlines the process of developing and deploying advanced machine learning solutions.