Alternatives to Grounded Language Model (GLM)

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.

Atla's Selene 1 API delivers cutting-edge AI evaluation models, empowering developers to set personalized assessment standards and achieve precise evaluations of their AI applications' effectiveness. Selene surpasses leading models on widely recognized evaluation benchmarks, guaranteeing trustworthy and accurate assessments. Users benefit from the ability to tailor evaluations to their unique requirements via the Alignment Platform, which supports detailed analysis and customized scoring systems. This API not only offers actionable feedback along with precise evaluation scores but also integrates smoothly into current workflows. It features established metrics like relevance, correctness, helpfulness, faithfulness, logical coherence, and conciseness, designed to tackle prevalent evaluation challenges, such as identifying hallucinations in retrieval-augmented generation scenarios or contrasting results with established ground truth data. Furthermore, the flexibility of the API allows developers to innovate and refine their evaluation methods continuously, making it an invaluable tool for enhancing AI application performance.

We present the Granite series of decoder-only code models specifically designed for tasks involving code generation, such as debugging, code explanation, and documentation, utilizing programming languages across a spectrum of 116 different types. An extensive assessment of the Granite Code model family across various tasks reveals that these models consistently achieve leading performance compared to other open-source code language models available today. Among the notable strengths of Granite Code models are: Versatile Code LLM: The Granite Code models deliver competitive or top-tier results across a wide array of code-related tasks, which include code generation, explanation, debugging, editing, translation, and beyond, showcasing their capacity to handle various coding challenges effectively. Additionally, their adaptability makes them suitable for both simple and complex coding scenarios. Reliable Enterprise-Grade LLM: All models in this series are developed using data that complies with licensing requirements and is gathered in alignment with IBM's AI Ethics guidelines, ensuring trustworthy usage for enterprise applications.

An advanced End-to-End MLLM is designed to accept various forms of references and effectively ground responses. The Ferret Model utilizes a combination of Hybrid Region Representation and a Spatial-aware Visual Sampler, which allows for detailed and flexible referring and grounding capabilities within the MLLM framework. The GRIT Dataset, comprising approximately 1.1 million entries, serves as a large-scale and hierarchical dataset specifically crafted for robust instruction tuning in the ground-and-refer category. Additionally, the Ferret-Bench is a comprehensive multimodal evaluation benchmark that simultaneously assesses referring, grounding, semantics, knowledge, and reasoning, ensuring a well-rounded evaluation of the model's capabilities. This intricate setup aims to enhance the interaction between language and visual data, paving the way for more intuitive AI systems.

Amazon Titan consists of a collection of sophisticated foundation models from AWS, aimed at boosting generative AI applications with exceptional performance and adaptability. Leveraging AWS's extensive expertise in AI and machine learning developed over 25 years, Titan models cater to various applications, including text generation, summarization, semantic search, and image creation. These models prioritize responsible AI practices by integrating safety features and fine-tuning options. Additionally, they allow for customization using your data through Retrieval Augmented Generation (RAG), which enhances accuracy and relevance, thus making them suitable for a wide array of both general and specialized AI tasks. With their innovative design and robust capabilities, Titan models represent a significant advancement in the field of artificial intelligence.

Kimi K2 represents a cutting-edge series of open-source large language models utilizing a mixture-of-experts (MoE) architecture, with a staggering 1 trillion parameters in total and 32 billion activated parameters tailored for optimized task execution. Utilizing the Muon optimizer, it has been trained on a substantial dataset of over 15.5 trillion tokens, with its performance enhanced by MuonClip’s attention-logit clamping mechanism, resulting in remarkable capabilities in areas such as advanced knowledge comprehension, logical reasoning, mathematics, programming, and various agentic operations. Moonshot AI offers two distinct versions: Kimi-K2-Base, designed for research-level fine-tuning, and Kimi-K2-Instruct, which is pre-trained for immediate applications in chat and tool interactions, facilitating both customized development and seamless integration of agentic features. Comparative benchmarks indicate that Kimi K2 surpasses other leading open-source models and competes effectively with top proprietary systems, particularly excelling in coding and intricate task analysis. Furthermore, it boasts a generous context length of 128 K tokens, compatibility with tool-calling APIs, and support for industry-standard inference engines, making it a versatile option for various applications. The innovative design and features of Kimi K2 position it as a significant advancement in the field of artificial intelligence language processing.

Cohere has launched Command A, an advanced AI model engineered to enhance efficiency while using minimal computational resources. This model not only competes with but also surpasses other leading models such as GPT-4 and DeepSeek-V3 in various enterprise tasks that require agentic capabilities, all while dramatically lowering computing expenses. Command A is specifically designed for applications that demand rapid and efficient AI solutions, enabling organizations to carry out complex tasks across multiple fields without compromising on performance or computational efficiency. Its innovative architecture allows businesses to harness the power of AI effectively, streamlining operations and driving productivity.

The outputs generated by Command’s model are accompanied by precise citations that help reduce the chances of misinformation while providing additional context drawn from the original sources. Command is capable of creating product descriptions, assisting in email composition, proposing sample press releases, and much more. You can engage Command with multiple inquiries about a document to categorize it, retrieve specific information, or address general questions pertaining to the content. While answering a handful of questions about a single document can save valuable time, applying this process to thousands of documents can lead to significant time savings for a business. This suite of scalable models achieves a remarkable balance between high efficiency and robust accuracy, empowering organizations to transition from experimental stages to fully operational AI solutions. By leveraging these capabilities, companies can enhance their productivity and streamline their workflows effectively.

Falcon Mamba 7B marks a significant milestone as the inaugural open-source State Space Language Model (SSLM), presenting a revolutionary architecture within the Falcon model family. Celebrated as the premier open-source SSLM globally by Hugging Face, it establishes a new standard for efficiency in artificial intelligence. In contrast to conventional transformers, SSLMs require significantly less memory and can produce lengthy text sequences seamlessly without extra resource demands. Falcon Mamba 7B outperforms top transformer models, such as Meta’s Llama 3.1 8B and Mistral’s 7B, demonstrating enhanced capabilities. This breakthrough not only highlights Abu Dhabi’s dedication to pushing the boundaries of AI research but also positions the region as a pivotal player in the global AI landscape. Such advancements are vital for fostering innovation and collaboration in technology.

PaLM 2 represents the latest evolution in large language models, continuing Google's tradition of pioneering advancements in machine learning and ethical AI practices. It demonstrates exceptional capabilities in complex reasoning activities such as coding, mathematics, classification, answering questions, translation across languages, and generating natural language, surpassing the performance of previous models, including its predecessor PaLM. This enhanced performance is attributed to its innovative construction, which combines optimal computing scalability, a refined mixture of datasets, and enhancements in model architecture. Furthermore, PaLM 2 aligns with Google's commitment to responsible AI development and deployment, having undergone extensive assessments to identify potential harms, biases, and practical applications in both research and commercial products. This model serves as a foundation for other cutting-edge applications, including Med-PaLM 2 and Sec-PaLM, while also powering advanced AI features and tools at Google, such as Bard and the PaLM API. Additionally, its versatility makes it a significant asset in various fields, showcasing the potential of AI to enhance productivity and innovation.

OpenGPT-X is an initiative based in Germany that is dedicated to creating large AI language models specifically designed to meet the needs of Europe, highlighting attributes such as adaptability, reliability, multilingual support, and open-source accessibility. This initiative unites various partners to encompass the full spectrum of the generative AI value chain, which includes scalable, GPU-powered infrastructure and data for training expansive language models, alongside model design and practical applications through prototypes and proofs of concept. The primary goal of OpenGPT-X is to promote innovative research with a significant emphasis on business applications, thus facilitating the quicker integration of generative AI within the German economic landscape. Additionally, the project places a strong importance on the ethical development of AI, ensuring that the models developed are both reliable and consistent with European values and regulations. Furthermore, OpenGPT-X offers valuable resources such as the LLM Workbook and a comprehensive three-part reference guide filled with examples and resources to aid users in grasping the essential features of large AI language models, ultimately fostering a deeper understanding of this technology. By providing these tools, OpenGPT-X not only supports the technical development of AI but also encourages responsible usage and implementation across various sectors.

The Gemma family consists of advanced, lightweight models developed using the same innovative research and technology as the Gemini models. These cutting-edge models are equipped with robust security features that promote responsible and trustworthy AI applications, achieved through carefully curated data sets and thorough refinements. Notably, Gemma models excel in their various sizes—2B, 7B, 9B, and 27B—often exceeding the performance of some larger open models. With the introduction of Keras 3.0, users can experience effortless integration with JAX, TensorFlow, and PyTorch, providing flexibility in framework selection based on specific tasks. Designed for peak performance and remarkable efficiency, Gemma 2 is specifically optimized for rapid inference across a range of hardware platforms. Furthermore, the Gemma family includes diverse models that cater to distinct use cases, ensuring they adapt effectively to user requirements. These lightweight language models feature a decoder and have been trained on an extensive array of textual data, programming code, and mathematical concepts, which enhances their versatility and utility in various applications.

The MiniMax‑M1 model, introduced by MiniMax AI and licensed under Apache 2.0, represents a significant advancement in hybrid-attention reasoning architecture. With an extraordinary capacity for handling a 1 million-token context window and generating outputs of up to 80,000 tokens, it facilitates in-depth analysis of lengthy texts. Utilizing a cutting-edge CISPO algorithm, MiniMax‑M1 was trained through extensive reinforcement learning, achieving completion on 512 H800 GPUs in approximately three weeks. This model sets a new benchmark in performance across various domains, including mathematics, programming, software development, tool utilization, and understanding of long contexts, either matching or surpassing the capabilities of leading models in the field. Additionally, users can choose between two distinct variants of the model, each with a thinking budget of either 40K or 80K, and access the model's weights and deployment instructions on platforms like GitHub and Hugging Face. Such features make MiniMax‑M1 a versatile tool for developers and researchers alike.

A marketplace driven by community engagement allows developers to create specialized expert LLMs focused on technical subjects, surpassing traditional AI models in performance. These Lunes significantly reduce inaccuracies in technical inquiries by continuously updating themselves with information from a variety of technical knowledge sources, including GitHub repositories and official documentation. Users can receive references akin to those provided by Perplexity, and access numerous Lunes built by other users, which range from those trained on open-source tools to well-curated collections of technology blog articles. You can also develop your own Lune by aggregating resources, including personal projects, to gain visibility. Our API seamlessly integrates with OpenAI’s, facilitating easy compatibility with tools like Cursor, Continue, and other applications that utilize OpenAI-compatible models. Conversations can effortlessly transition from your IDE to Lune Web at any point, enhancing user experience. Contributions made during chats can earn you compensation for every piece of feedback that gets approved. Alternatively, you can create a public Lune and share it widely, earning money based on its popularity and user engagement. This innovative approach not only fosters collaboration but also rewards users for their expertise and creativity.

LearnLM is a novel, experimental model tailored for specific tasks, developed to align with the principles of learning science for enhanced teaching and learning experiences. It is adept at following system prompts such as "You are an expert tutor," and promotes active engagement in learning by facilitating practice and offering timely feedback. By effectively managing cognitive load, the model delivers pertinent and well-organized information through various modalities, while also adjusting to the individual learner’s objectives and requirements, grounding its responses in suitable resources. Furthermore, LearnLM encourages curiosity, sustaining learner motivation throughout their educational pursuits, and fosters metacognitive skills by assisting learners in planning, monitoring, and reflecting on their academic progress. This groundbreaking model is currently accessible for experimentation within AI Studio, allowing educators and researchers to explore its potential in real-world applications. Ultimately, LearnLM represents a significant step forward in the integration of AI within educational contexts.

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.

Llama 4 Behemoth, with 288 billion active parameters, is Meta's flagship AI model, setting new standards for multimodal performance. Outpacing its predecessors like GPT-4.5 and Claude Sonnet 3.7, it leads the field in STEM benchmarks, offering cutting-edge results in tasks such as problem-solving and reasoning. Designed as the teacher model for the Llama 4 series, Behemoth drives significant improvements in model quality and efficiency through distillation. Although still in development, Llama 4 Behemoth is shaping the future of AI with its unparalleled intelligence, particularly in math, image, and multilingual tasks.

GPT-5 Thinking is a specialized reasoning component of the GPT-5 platform that activates when queries require deeper thought and complex problem-solving. Unlike the quick-response GPT-5 base model, GPT-5 Thinking carefully processes multifaceted questions, delivering richer and more precise answers. This enhanced reasoning mode excels in reducing factual errors and hallucinations by analyzing information more thoroughly and applying multi-step logic. It also improves transparency by clearly stating when certain tasks cannot be completed due to missing data or unsupported requests. Safety is a core focus, with GPT-5 Thinking trained to balance helpfulness and risk, especially in sensitive or dual-use scenarios. The model seamlessly switches between fast and deep thinking based on conversation complexity and user intent. With improved instruction following and reduced sycophancy, GPT-5 Thinking offers more natural, confident, and thoughtful interactions. It is accessible to all users as part of GPT-5’s unified system, enhancing both everyday productivity and expert applications.

ERNIE X1.1 is Baidu’s latest reasoning AI model, designed to raise the bar for accuracy, reliability, and action-oriented intelligence. Compared to ERNIE X1, it delivers a 34.8% boost in factual accuracy, a 12.5% improvement in instruction compliance, and a 9.6% gain in agentic behavior. Benchmarks show that it outperforms DeepSeek R1-0528 and matches the capabilities of advanced models such as GPT-5 and Gemini 2.5 Pro. The model builds upon ERNIE 4.5 with additional mid-training and post-training phases, reinforced by end-to-end reinforcement learning. This approach helps minimize hallucinations while ensuring closer alignment to user intent. The agentic upgrades allow it to plan, make decisions, and execute tasks more effectively than before. Users can access ERNIE X1.1 through ERNIE Bot, Wenxiaoyan, or via API on Baidu’s Qianfan platform. Altogether, the model delivers stronger reasoning capabilities for developers and enterprises that demand high-performance AI.

Anthropic’s release of Claude Opus 4.5 introduces a frontier AI model that excels at coding, complex reasoning, deep research, and long-context tasks. It sets new performance records on real-world engineering benchmarks, handling multi-system debugging, ambiguous instructions, and cross-domain problem solving with greater precision than earlier versions. Testers and early customers reported that Opus 4.5 “just gets it,” offering creative reasoning strategies that even benchmarks fail to anticipate. Beyond raw capability, the model brings stronger alignment and safety, with notable advances in prompt-injection resistance and behavior consistency in high-stakes scenarios. The Claude Developer Platform also gains richer controls including effort tuning, multi-agent orchestration, and context management improvements that significantly boost efficiency. Claude Code becomes more powerful with enhanced planning abilities, multi-session desktop support, and better execution of complex development workflows. In the Claude apps, extended memory and automatic context summarization enable longer, uninterrupted conversations. Together, these upgrades showcase Opus 4.5 as a highly capable, secure, and versatile model designed for both professional workloads and everyday use.

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.

Jamba stands out as the most potent and effective long context model, specifically designed for builders while catering to enterprise needs. With superior latency compared to other leading models of similar sizes, Jamba boasts a remarkable 256k context window, the longest that is openly accessible. Its innovative Mamba-Transformer MoE architecture focuses on maximizing cost-effectiveness and efficiency. Key features available out of the box include function calls, JSON mode output, document objects, and citation mode, all designed to enhance user experience. Jamba 1.5 models deliver exceptional performance throughout their extensive context window and consistently achieve high scores on various quality benchmarks. Enterprises can benefit from secure deployment options tailored to their unique requirements, allowing for seamless integration into existing systems. Jamba can be easily accessed on our robust SaaS platform, while deployment options extend to strategic partners, ensuring flexibility for users. For organizations with specialized needs, we provide dedicated management and continuous pre-training, ensuring that every client can leverage Jamba’s capabilities to the fullest. This adaptability makes Jamba a prime choice for enterprises looking for cutting-edge solutions.

Mistral Medium 3 is an innovative AI model designed to offer high performance at a significantly lower cost, making it an attractive solution for enterprises. It integrates seamlessly with both on-premises and cloud environments, supporting hybrid deployments for more flexibility. This model stands out in professional use cases such as coding, STEM tasks, and multimodal understanding, where it achieves near-competitive results against larger, more expensive models. Additionally, Mistral Medium 3 allows businesses to deploy custom post-training and integrate it into existing systems, making it adaptable to various industry needs. With its impressive performance in coding tasks and real-world human evaluations, Mistral Medium 3 is a cost-effective solution that enables companies to implement AI into their workflows. Its enterprise-focused features, including continuous pretraining and domain-specific fine-tuning, make it a reliable tool for sectors like healthcare, financial services, and energy.

Grok 4 Heavy represents xAI’s flagship AI model, leveraging a multi-agent architecture to deliver exceptional reasoning, problem-solving, and multimodal understanding. Developed using the Colossus supercomputer, it achieves a remarkable 50% score on the HLE benchmark, placing it among the leading AI models worldwide. This version can process text, images, and is expected to soon support video inputs, enabling richer contextual comprehension. Grok 4 Heavy is designed for advanced users, including developers and researchers, who demand state-of-the-art AI capabilities for complex scientific and technical tasks. Available exclusively through a $300/month SuperGrok Heavy subscription, it offers early access to future innovations like video generation. xAI has addressed past controversies by strengthening content moderation and removing harmful prompts. The platform aims to push AI boundaries while balancing ethical considerations. Grok 4 Heavy is positioned as a formidable competitor to other leading AI systems.

Amazon Nova Sonic is an advanced speech-to-speech model that offers real-time, lifelike voice interactions while maintaining exceptional price efficiency. By integrating speech comprehension and generation into one cohesive model, it allows developers to craft engaging and fluid conversational AI solutions with minimal delay. This system fine-tunes its replies by analyzing the prosody of the input speech, including elements like rhythm and tone, which leads to more authentic conversations. Additionally, Nova Sonic features function calling and agentic workflows that facilitate interactions with external services and APIs, utilizing knowledge grounding with enterprise data through Retrieval-Augmented Generation (RAG). Its powerful speech understanding capabilities encompass both American and British English across a variety of speaking styles and acoustic environments, with plans to incorporate more languages in the near future. Notably, Nova Sonic manages interruptions from users seamlessly while preserving the context of the conversation, demonstrating its resilience against background noise interference and enhancing the overall user experience. This technology represents a significant leap forward in conversational AI, ensuring that interactions are not only efficient but also genuinely engaging.

The Gemini 2.0 Flash AI model signifies a revolutionary leap in high-speed, intelligent computing, aiming to redefine standards in real-time language processing and decision-making capabilities. By enhancing the strong foundation laid by its predecessor, it features advanced neural architecture and significant optimization breakthroughs that facilitate quicker and more precise responses. Tailored for applications that demand immediate processing and flexibility, such as live virtual assistants, automated trading systems, and real-time analytics, Gemini 2.0 Flash excels in various contexts. Its streamlined and efficient design allows for effortless deployment across cloud, edge, and hybrid environments, making it adaptable to diverse technological landscapes. Furthermore, its superior contextual understanding and multitasking abilities equip it to manage complex and dynamic workflows with both accuracy and speed, solidifying its position as a powerful asset in the realm of artificial intelligence. With each iteration, technology continues to advance, and models like Gemini 2.0 Flash pave the way for future innovations in the field.

Palmyra represents a collection of Large Language Models (LLMs) specifically designed to deliver accurate and reliable outcomes in business settings. These models shine in various applications, including answering questions, analyzing images, and supporting more than 30 languages, with options for fine-tuning tailored to sectors such as healthcare and finance. Remarkably, the Palmyra models have secured top positions in notable benchmarks such as Stanford HELM and PubMedQA, with Palmyra-Fin being the first to successfully clear the CFA Level III examination. Writer emphasizes data security by refraining from utilizing client data for training or model adjustments, adhering to a strict zero data retention policy. The Palmyra suite features specialized models, including Palmyra X 004, which boasts tool-calling functionalities; Palmyra Med, created specifically for the healthcare industry; Palmyra Fin, focused on financial applications; and Palmyra Vision, which delivers sophisticated image and video processing capabilities. These advanced models are accessible via Writer's comprehensive generative AI platform, which incorporates graph-based Retrieval Augmented Generation (RAG) for enhanced functionality. With continual advancements and improvements, Palmyra aims to redefine the landscape of enterprise-level AI solutions.

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.

Tülu 3 is a cutting-edge language model created by the Allen Institute for AI (Ai2) that aims to improve proficiency in fields like knowledge, reasoning, mathematics, coding, and safety. It is based on the Llama 3 Base and undergoes a detailed four-stage post-training regimen: careful prompt curation and synthesis, supervised fine-tuning on a wide array of prompts and completions, preference tuning utilizing both off- and on-policy data, and a unique reinforcement learning strategy that enhances targeted skills through measurable rewards. Notably, this open-source model sets itself apart by ensuring complete transparency, offering access to its training data, code, and evaluation tools, thus bridging the performance divide between open and proprietary fine-tuning techniques. Performance assessments reveal that Tülu 3 surpasses other models with comparable sizes, like Llama 3.1-Instruct and Qwen2.5-Instruct, across an array of benchmarks, highlighting its effectiveness. The continuous development of Tülu 3 signifies the commitment to advancing AI capabilities while promoting an open and accessible approach to technology.

Gemini 3, the latest model from Google DeepMind, establishes a new standard for artificial intelligence by achieving cutting-edge reasoning capabilities and multimodal comprehension across various formats including text, images, and videos. It significantly outperforms its earlier version in critical AI assessments and showcases its strengths in intricate areas like scientific reasoning, advanced programming, spatial reasoning, and visual or video interpretation. The introduction of the innovative “Deep Think” mode takes performance to an even higher level, demonstrating superior reasoning abilities for exceptionally difficult tasks and surpassing the Gemini 3 Pro in evaluations such as Humanity’s Last Exam and ARC-AGI. Now accessible within Google’s ecosystem, Gemini 3 empowers users to engage in learning, developmental projects, and strategic planning with unprecedented sophistication. With context windows extending up to one million tokens and improved media-processing capabilities, along with tailored configurations for various tools, the model enhances precision, depth, and adaptability for practical applications, paving the way for more effective workflows across diverse industries. This advancement signals a transformative shift in how AI can be leveraged for real-world challenges.

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.

Mercury, the groundbreaking creation from Inception Labs, represents the first large language model at a commercial scale that utilizes diffusion technology, achieving a remarkable tenfold increase in processing speed while also lowering costs in comparison to standard autoregressive models. Designed for exceptional performance in reasoning, coding, and the generation of structured text, Mercury can handle over 1000 tokens per second when operating on NVIDIA H100 GPUs, positioning it as one of the most rapid LLMs on the market. In contrast to traditional models that produce text sequentially, Mercury enhances its responses through a coarse-to-fine diffusion strategy, which boosts precision and minimizes instances of hallucination. Additionally, with the inclusion of Mercury Coder, a tailored coding module, developers are empowered to take advantage of advanced AI-assisted code generation that boasts remarkable speed and effectiveness. This innovative approach not only transforms coding practices but also sets a new benchmark for the capabilities of AI in various applications.

DeepSeek-V2 is a cutting-edge Mixture-of-Experts (MoE) language model developed by DeepSeek-AI, noted for its cost-effective training and high-efficiency inference features. It boasts an impressive total of 236 billion parameters, with only 21 billion active for each token, and is capable of handling a context length of up to 128K tokens. The model utilizes advanced architectures such as Multi-head Latent Attention (MLA) to optimize inference by minimizing the Key-Value (KV) cache and DeepSeekMoE to enable economical training through sparse computations. Compared to its predecessor, DeepSeek 67B, this model shows remarkable improvements, achieving a 42.5% reduction in training expenses, a 93.3% decrease in KV cache size, and a 5.76-fold increase in generation throughput. Trained on an extensive corpus of 8.1 trillion tokens, DeepSeek-V2 demonstrates exceptional capabilities in language comprehension, programming, and reasoning tasks, positioning it as one of the leading open-source models available today. Its innovative approach not only elevates its performance but also sets new benchmarks within the field of artificial intelligence.

OpenAI’s GPT-5 Pro represents the pinnacle of AI reasoning power, offering enhanced capabilities for solving the toughest problems with unparalleled precision and depth. This version leverages extensive parallel compute resources to deliver highly accurate, detailed answers that outperform prior models across challenging scientific, medical, mathematical, and programming benchmarks. GPT-5 Pro is particularly effective in handling multi-step, complex queries that require sustained focus and logical reasoning. Experts consistently rate its outputs as more comprehensive, relevant, and error-resistant than those from standard GPT-5. It seamlessly integrates with existing ChatGPT offerings, allowing Pro users to access this powerful reasoning mode for demanding tasks. The model’s ability to dynamically allocate “thinking” resources ensures efficient and expert-level responses. Additionally, GPT-5 Pro features improved safety, reduced hallucinations, and better transparency about its capabilities and limitations. It empowers professionals and researchers to push the boundaries of what AI can achieve.

The overwhelming amount of information available poses a significant challenge to advancements in science. With the rapid expansion of scientific literature and data, pinpointing valuable insights within this vast sea of information has become increasingly difficult. Nowadays, people rely on search engines to access scientific knowledge, yet these tools alone cannot effectively categorize and organize this complex information. Galactica is an advanced language model designed to capture, synthesize, and analyze scientific knowledge. It is trained on a diverse array of scientific materials, including research papers, reference texts, knowledge databases, and other relevant resources. In various scientific tasks, Galactica demonstrates superior performance compared to existing models. For instance, on technical knowledge assessments involving LaTeX equations, Galactica achieves a score of 68.2%, significantly higher than the 49.0% of the latest GPT-3 model. Furthermore, Galactica excels in reasoning tasks, outperforming Chinchilla in mathematical MMLU with scores of 41.3% to 35.7%, and surpassing PaLM 540B in MATH with a notable 20.4% compared to 8.8%. This indicates that Galactica not only enhances accessibility to scientific information but also improves our ability to reason through complex scientific queries.

Llama (Large Language Model Meta AI) stands as a cutting-edge foundational large language model aimed at helping researchers push the boundaries of their work within this area of artificial intelligence. By providing smaller yet highly effective models like Llama, the research community can benefit even if they lack extensive infrastructure, thus promoting greater accessibility in this dynamic and rapidly evolving domain. Creating smaller foundational models such as Llama is advantageous in the landscape of large language models, as it demands significantly reduced computational power and resources, facilitating the testing of innovative methods, confirming existing research, and investigating new applications. These foundational models leverage extensive unlabeled datasets, making them exceptionally suitable for fine-tuning across a range of tasks. We are offering Llama in multiple sizes (7B, 13B, 33B, and 65B parameters), accompanied by a detailed Llama model card that outlines our development process while adhering to our commitment to Responsible AI principles. By making these resources available, we aim to empower a broader segment of the research community to engage with and contribute to advancements in AI.

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.

K2 Think represents a groundbreaking open-source advanced reasoning model that has been developed in collaboration between the Institute of Foundation Models at MBZUAI and G42. Even with its relatively modest 32 billion parameters, K2 Think achieves performance that rivals that of leading models with significantly larger parameter counts. Its strength lies in mathematical reasoning, where it has secured top rankings on prestigious benchmarks such as AIME ’24/’25, HMMT ’25, and OMNI-Math-HARD. This model is part of a wider initiative of UAE-developed open models, which includes Jais (for Arabic), NANDA (for Hindi), and SHERKALA (for Kazakh), and it builds upon the groundwork established by the K2-65B, a fully reproducible open-source foundation model released in 2024. K2 Think is crafted to be open, efficient, and adaptable, featuring a web app interface that facilitates user exploration, and its innovative approach to parameter positioning marks a significant advancement in the realm of compact architectures for high-level AI reasoning. Additionally, its development highlights a commitment to enhancing access to state-of-the-art AI technologies in various languages and domains.

Claude Sonnet 4.5 represents Anthropic's latest advancement in AI, crafted to thrive in extended coding environments, complex workflows, and heavy computational tasks while prioritizing safety and alignment. It sets new benchmarks with its top-tier performance on the SWE-bench Verified benchmark for software engineering and excels in the OSWorld benchmark for computer usage, demonstrating an impressive capacity to maintain concentration for over 30 hours on intricate, multi-step assignments. Enhancements in tool management, memory capabilities, and context interpretation empower the model to engage in more advanced reasoning, leading to a better grasp of various fields, including finance, law, and STEM, as well as a deeper understanding of coding intricacies. The system incorporates features for context editing and memory management, facilitating prolonged dialogues or multi-agent collaborations, while it also permits code execution and the generation of files within Claude applications. Deployed at AI Safety Level 3 (ASL-3), Sonnet 4.5 is equipped with classifiers that guard against inputs or outputs related to hazardous domains and includes defenses against prompt injection, ensuring a more secure interaction. This model signifies a significant leap forward in the intelligent automation of complex tasks, aiming to reshape how users engage with AI technologies.

Chinchilla is an advanced language model that operates with a compute budget comparable to Gopher while having 70 billion parameters and utilizing four times the amount of data. This model consistently and significantly surpasses Gopher (280 billion parameters), as well as GPT-3 (175 billion), Jurassic-1 (178 billion), and Megatron-Turing NLG (530 billion), across a wide variety of evaluation tasks. Additionally, Chinchilla's design allows it to use significantly less computational power during the fine-tuning and inference processes, which greatly enhances its applicability in real-world scenarios. Notably, Chinchilla achieves a remarkable average accuracy of 67.5% on the MMLU benchmark, marking over a 7% enhancement compared to Gopher, showcasing its superior performance in the field. This impressive capability positions Chinchilla as a leading contender in the realm of language models.

Mistral AI has unveiled two cutting-edge models specifically designed for on-device computing and edge use cases, collectively referred to as "les Ministraux": Ministral 3B and Ministral 8B. These innovative models stand out due to their capabilities in knowledge retention, commonsense reasoning, function-calling, and overall efficiency, all while remaining within the sub-10B parameter range. They boast support for a context length of up to 128k, making them suitable for a diverse range of applications such as on-device translation, offline smart assistants, local analytics, and autonomous robotics. Notably, Ministral 8B incorporates an interleaved sliding-window attention mechanism, which enhances both the speed and memory efficiency of inference processes. Both models are adept at serving as intermediaries in complex multi-step workflows, skillfully managing functions like input parsing, task routing, and API interactions based on user intent, all while minimizing latency and operational costs. Benchmark results reveal that les Ministraux consistently exceed the performance of similar models across a variety of tasks, solidifying their position in the market. As of October 16, 2024, these models are now available for developers and businesses, with Ministral 8B being offered at a competitive rate of $0.1 for every million tokens utilized. This pricing structure enhances accessibility for users looking to integrate advanced AI capabilities into their solutions.

Llama 4 Scout is an advanced multimodal AI model with 17 billion active parameters, offering industry-leading performance with a 10 million token context length. This enables it to handle complex tasks like multi-document summarization and detailed code reasoning with impressive accuracy. Scout surpasses previous Llama models in both text and image understanding, making it an excellent choice for applications that require a combination of language processing and image analysis. Its powerful capabilities in long-context tasks and image-grounding applications set it apart from other models in its class, providing superior results for a wide range of industries.

LFM-3B offers outstanding performance relative to its compact size, securing its top position among models with 3 billion parameters, hybrids, and RNNs, while surpassing earlier generations of 7 billion and 13 billion parameter models. In addition, it matches the performance of Phi-3.5-mini across several benchmarks, all while being 18.4% smaller in size. This makes LFM-3B the perfect option for mobile applications and other edge-based text processing needs, illustrating its versatility and efficiency in a variety of settings.

We are excited to introduce Jurassic-2, the newest iteration of AI21 Studio's foundation models, which represents a major advancement in artificial intelligence, boasting exceptional quality and innovative features. In addition to this, we are unveiling our tailored APIs that offer seamless reading and writing functionalities, surpassing those of our rivals. At AI21 Studio, our mission is to empower developers and businesses to harness the potential of reading and writing AI, facilitating the creation of impactful real-world applications. Today signifies a pivotal moment with the launch of Jurassic-2 and our Task-Specific APIs, enabling you to effectively implement generative AI in production settings. Known informally as J2, Jurassic-2 showcases remarkable enhancements in quality, including advanced zero-shot instruction-following, minimized latency, and support for multiple languages. Furthermore, our specialized APIs are designed to provide developers with top-tier tools that excel in executing specific reading and writing tasks effortlessly, ensuring you have everything needed to succeed in your projects. Together, these advancements set a new standard in the AI landscape, paving the way for innovative solutions.

Tencent's Hunyuan-TurboS represents a cutting-edge AI model crafted to deliver swift answers and exceptional capabilities across multiple fields, including knowledge acquisition, mathematical reasoning, and creative endeavors. Departing from earlier models that relied on "slow thinking," this innovative system significantly boosts response rates, achieving a twofold increase in word output speed and cutting down first-word latency by 44%. With its state-of-the-art architecture, Hunyuan-TurboS not only enhances performance but also reduces deployment expenses. The model skillfully integrates fast thinking—prompt, intuition-driven responses—with slow thinking—methodical logical analysis—ensuring timely and precise solutions in a wide array of situations. Its remarkable abilities are showcased in various benchmarks, positioning it competitively alongside other top AI models such as GPT-4 and DeepSeek V3, thus marking a significant advancement in AI performance. As a result, Hunyuan-TurboS is poised to redefine expectations in the realm of artificial intelligence applications.