Best Foundation Models in New Zealand - Page 9

Qwen3.6-Plus is a state-of-the-art AI model designed to support real-world agentic applications, advanced coding, and multimodal reasoning. Developed by the Qwen team under Alibaba Cloud, it offers a significant upgrade over previous versions with improved performance across coding, reasoning, and tool usage tasks. The model features a 1 million token context window, enabling it to handle long and complex workflows with high accuracy. It excels in agentic coding scenarios, including debugging, repository-level problem solving, and automated development tasks. Qwen3.6-Plus integrates reasoning, memory, and execution into a unified system, allowing it to operate as a highly capable autonomous agent. Its multimodal capabilities enable it to process and analyze text, images, videos, and documents for deeper insights. The model supports real-time tool usage and long-horizon planning, making it ideal for enterprise and developer use cases. It is accessible via API through Alibaba Cloud Model Studio and integrates with popular coding tools and assistants. Developers can leverage features like preserved reasoning context to improve performance in multi-step tasks. Overall, Qwen3.6-Plus empowers businesses and developers to build intelligent, scalable, and autonomous AI-driven applications.

Xiaomi MiMo-V2.5-Pro is a next-generation open-source AI model designed for advanced reasoning, coding, and long-horizon task execution. It uses a Mixture-of-Experts architecture with over one trillion parameters and a large active parameter set for efficient performance. The model supports an extended context window of up to one million tokens, allowing it to handle complex, multi-step workflows. It is built to perform autonomous tasks, including software development, system design, and engineering optimization. Benchmark results show strong performance across coding, reasoning, and agent-based evaluation tests. MiMo-V2.5-Pro incorporates hybrid attention mechanisms to improve efficiency while maintaining accuracy across long contexts. It is optimized for token efficiency, reducing the computational cost of running complex tasks. The model can integrate with development tools and frameworks to support real-world applications. It is designed to complete tasks that would typically require significant human effort over extended periods. Xiaomi has made the model open source, enabling developers to access and customize it. By combining performance, scalability, and efficiency, MiMo-V2.5-Pro pushes the boundaries of modern AI capabilities.

Xiaomi MiMo-V2.5 is a next-generation open-source AI model that combines agentic intelligence with multimodal capabilities. It is designed to process and understand text, images, and audio within a single architecture. The model uses a sparse Mixture-of-Experts framework with a large parameter count to deliver efficient and scalable performance. It supports a context window of up to one million tokens, allowing it to handle long and complex workflows. MiMo-V2.5 integrates visual and audio encoders to improve perception and cross-modal reasoning. It is capable of performing tasks such as coding, reasoning, and multimodal analysis with strong accuracy. Benchmark results show competitive performance compared to leading AI models in both agentic and multimodal tasks. The model is optimized for token efficiency, balancing performance with lower computational cost. It is designed for real-world applications that require both reasoning and perception. Xiaomi has open-sourced the model, making it accessible for developers and researchers. By combining multimodality, scalability, and efficiency, MiMo-V2.5 pushes forward the development of advanced AI systems.

ERNIE 5.1 is Baidu’s next-generation large language model engineered to provide advanced reasoning, autonomous agent capabilities, creative writing performance, and enterprise-grade AI intelligence with highly optimized efficiency. Built on the pre-training foundation of ERNIE 5.0, the model significantly reduces parameter size and computational requirements while still delivering leading performance across major international AI benchmarks. ERNIE 5.1 demonstrates strong capabilities in reasoning, mathematical problem solving, knowledge retrieval, search tasks, and agentic workflows that allow it to handle complex multi-step operations and decision-making scenarios. The platform introduces a fully asynchronous reinforcement learning architecture designed to improve scalability, training efficiency, resource utilization, and long-horizon task stability for large-scale AI development. Baidu also implemented a multi-stage reinforcement learning pipeline that separates expert capability training from unified capability fusion, allowing the model to specialize in areas such as coding, reasoning, search, and conversational intelligence without creating performance conflicts between domains. ERNIE 5.1 supports advanced creative generation with improved emotional understanding, narrative structure control, stylistic adaptability, and contextual awareness for writing-intensive applications. The model performs competitively against leading closed-source global AI systems in knowledge benchmarks, reasoning evaluations, and creative content generation tasks. ERNIE 5.1 is also integrated into creative production platforms, AI storytelling systems, roleplay applications, and agentic AI environments that support content creators and enterprise workflows.

Seed2.1 Turbo represents an advanced AI productivity model that is adept at tackling intricate real-world challenges through its robust general-agent capabilities, coding proficiency, and multimodal functionality. Unlike traditional models that offer singular solutions, it is equipped to manage multi-step workflows aimed at achieving specific objectives, generating practical and actionable results across various tools and environments. In both professional settings and everyday tasks, it can assist with project management, document handling, data analysis, solution development, content organization, tool utilization, and synthesizing results. Additionally, it excels in educational, office, and research contexts, facilitating tasks such as crafting lesson-plan presentations, dissecting detailed spreadsheets, and generating comprehensive industry analyses. In the realm of software engineering, Seed2.1 Turbo facilitates complete project delivery, encompassing requirement analysis, feature development, bug resolution, environment configuration, terminal commands, and validation of outcomes, while also possessing a deep understanding of codebase structure, dependencies, and business logic to efficiently manage modifications. This model’s versatility makes it a valuable asset across a wide range of applications, ensuring that users can leverage AI to enhance productivity and streamline their workflows.

Qwen3.8-2.4T-A95B stands out as the most extensive open model within the Qwen3.8 series, offering advanced Qwen-Max-class features in a publicly accessible format. Constructed upon the solid framework of Qwen3.5, this model significantly enhances performance in areas such as coding, professional tasks, research, and complex, prolonged agentic activities, emphasizing the reliability of executing intricate, multi-step workflows to completion. Utilizing a cutting-edge mixture-of-experts architecture, it boasts an impressive total of 2.4 trillion parameters, with 95 billion of those being activated, featuring 512 experts and engaging 10 routed along with one shared expert simultaneously. The model accommodates a native context length of 262,144 tokens, which can be extended to around 1.01 million tokens, thereby providing substantial flexibility for various applications. Furthermore, improvements in agent execution, such as enhanced autonomous planning and better responsiveness to environmental feedback, contribute to its efficiency, while its broader compatibility with widely used agent frameworks and development tools facilitates seamless integration into existing systems, making it a versatile choice for developers and researchers alike.

Hy4 preview represents a cutting-edge open source Mixture-of-Experts flagship model tailored for a variety of real-world productivity tasks, including software engineering, office activities, game development, and scientific exploration. This model boasts a staggering total of 770 billion parameters, with 49 billion activated per token, and features an impressive 1 million-token context window, allowing it to efficiently manage large codebases, vast document collections, and complex multi-step processes. The architecture consists of 78 layers that integrate Gated DeepSeek Sparse Attention alongside IndexCache for reusing sparse indices across layers, while also employing identity Hyper-Connections to enhance the flow of information between layers. Additionally, a dedicated Multi-Token Prediction layer facilitates speculative decoding, further enhancing its capabilities. Hy4 preview is crafted to comprehend, plan, debug, and validate intricate engineering projects, while also achieving notable improvements in the quality of front-end visuals and interaction design, thereby making it an invaluable asset for professionals across various domains.

TimesFM-3 represents an advanced time series foundation model that excels in highly precise multivariate forecasting with a single forward pass. This model, which consists of 330 million parameters, has undergone pre-training on a vast corpus of real-world and synthetic time series data, totaling over 1 trillion time points, thereby enhancing the effectiveness and zero-shot generalization capabilities seen in previous TimesFM iterations. It is adept at simultaneously predicting numerous coevolving time series and understanding dependencies that bolster accuracy without the need for task-specific fine-tuning. Furthermore, it accommodates multiple forecasting targets, including both point and quantile predictions, and incorporates past covariates that are only available historically, alongside dynamic covariates that pertain to future events such as planned promotions, holidays, or weather changes. Utilizing a decoder-only transformer architecture, TimesFM-3 processes sequential data in segments of 32 time steps, employing alternating causal temporal attention and full variate attention to integrate patterns across both time and interrelated series effectively. As a result, it provides a robust tool for forecasting complex time-dependent phenomena in various applications.

Gemini 3.8 Flash stands out as Google's most advanced model for Flash, offering substantial enhancements compared to version 3.7 in areas such as software engineering, agent-based tasks, and intricate multi-step reasoning within specialized fields. Designed for extended coding projects and autonomous agents, it adeptly addresses complex engineering challenges in a comprehensive manner, ensuring the reliability essential for critical enterprise autonomy in specialized knowledge areas. This model excels particularly in quantitative and professional disciplines that demand sophisticated analysis and reporting, as well as in multi-step reasoning tasks spanning STEM, humanities, and professional domains. The improvements it showcases arise from a fundamental design decision: Gemini 3.8 Flash intensifies its focus on challenging tasks by conducting additional reasoning steps and utilizing tools iteratively, thus optimizing its performance. When operating at higher effort levels, it may consume more tokens to achieve superior outcomes, while developers also have the option to adjust to lower effort levels for varied results. Overall, this flexibility allows for tailored use based on project needs and desired outcomes.

Gemini 3.8 Flash Cyber represents Google's most advanced cybersecurity model, offering top-tier performance in identifying vulnerabilities and automating patching processes with remarkable speed for rapid iteration. Tailored for trusted defenders, it is accessible via the Fairwind Program. On CyberGym, a recognized industry benchmark for detecting vulnerabilities, this model showcases exceptional autonomous vulnerability discovery, outperforming both Gemini 3.5 Flash Cyber and larger frontier models. Furthermore, Google assessed its effectiveness on an internal benchmark that spans complex codebases across 20 programming languages, achieving a success rate of over 70% in identifying various vulnerabilities. Unlike many models that focus on offensive strategies, Gemini 3.8 Flash Cyber emphasizes the importance of fixing vulnerabilities, providing defenders with advanced tools that enhance their ability to stay ahead of cyber attackers. This focus on proactive defense represents a crucial shift in the cybersecurity landscape, prioritizing the safeguarding of systems over mere exploitation capabilities.

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.

The o3-mini by OpenAI is a streamlined iteration of the sophisticated o3 AI model, delivering robust reasoning skills in a more compact and user-friendly format. It specializes in simplifying intricate instructions into digestible steps, making it particularly adept at coding, competitive programming, and tackling mathematical and scientific challenges. This smaller model maintains the same level of accuracy and logical reasoning as the larger version, while operating with lower computational demands, which is particularly advantageous in environments with limited resources. Furthermore, o3-mini incorporates inherent deliberative alignment, promoting safe, ethical, and context-sensitive decision-making. Its versatility makes it an invaluable resource for developers, researchers, and enterprises striving for an optimal mix of performance and efficiency in their projects. The combination of these features positions o3-mini as a significant tool in the evolving landscape of AI-driven solutions.

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.

The o4-mini model, a more compact and efficient iteration of the o3 model, was developed to enhance reasoning capabilities and streamline performance. It excels in tasks requiring complex problem-solving, making it an ideal solution for users demanding more powerful AI. By refining its design, OpenAI has made significant strides in creating a model that balances efficiency with advanced capabilities. With this release, the o4-mini is poised to meet the growing need for smarter AI tools while maintaining the robust functionality of its predecessor. It plays a critical role in OpenAI’s ongoing efforts to push the boundaries of artificial intelligence ahead of the GPT-5 launch.

Microsoft Foundry Models centralizes more than 11,000 leading AI models, offering enterprises a single place to explore, compare, fine-tune, and deploy AI for any use case. It includes top-performing models from OpenAI, Anthropic, Cohere, Meta, Mistral AI, DeepSeek, Black Forest Labs, and Microsoft’s own Azure OpenAI offerings. Teams can search by task—such as reasoning, generation, multimodal, or domain-specific workloads—and instantly test models in a built-in playground. Foundry Models simplifies customization with ready-to-use fine-tuning pipelines that require no infrastructure setup. Developers can upload internal datasets to benchmark and evaluate model accuracy, ensuring the right fit for production environments. With seamless deployment into managed instances, organizations get automatic scaling, traffic management, and secure hosting. The platform is backed by Azure’s enterprise-grade security and over 100 compliance certifications, supporting regulated industries and global operations. By integrating discovery, testing, tuning, and deployment, Foundry Models dramatically shortens AI development cycles and speeds time to value.

Grok 4.1, developed by Elon Musk’s xAI, represents a major step forward in multimodal artificial intelligence. Built on the Colossus supercomputer, it supports input from text, images, and soon video—offering a more complete understanding of real-world data. This version significantly improves reasoning precision, enabling Grok to solve complex problems in science, engineering, and language with remarkable clarity. Developers and researchers can leverage Grok 4.1’s advanced APIs to perform deep contextual analysis, creative generation, and data-driven research. Its refined architecture allows it to outperform leading models in visual problem-solving and structured reasoning benchmarks. xAI has also strengthened the model’s moderation framework, addressing bias and ensuring more balanced responses. With its multimodal flexibility and intelligent output control, Grok 4.1 bridges the gap between analytical computation and human intuition. It’s a model designed not just to answer questions, but to understand and reason through them.

Nova 2 Omni is an innovative model that seamlessly integrates multimodal reasoning and generation, allowing it to comprehend and generate diverse types of content, including text, images, video, and audio. Its capability to process exceptionally large inputs, which can encompass hundreds of thousands of words or several hours of audiovisual material, enables it to maintain a coherent analysis across various formats. As a result, it can simultaneously analyze comprehensive product catalogs, extensive documents, customer reviews, and entire video libraries, providing teams with a singular system that eliminates the necessity for multiple specialized models. By managing mixed media within a unified workflow, Nova 2 Omni paves the way for new opportunities in both creative and operational automation. For instance, a marketing team can input product specifications, brand standards, reference visuals, and video content to effortlessly generate an entire campaign that includes messaging, social media content, and visuals, all in one streamlined process. This efficiency not only enhances productivity but also fosters innovation in how teams approach their marketing strategies.

GPT-5.4 is a next-generation AI model created by OpenAI to assist professionals with advanced knowledge work and software development tasks. It brings together major improvements in reasoning, coding, and automated workflows to deliver more capable and reliable results. The model can analyze large datasets, generate detailed reports, create presentations, and assist with spreadsheet modeling. GPT-5.4 also supports complex coding tasks and can help developers build, test, and debug software more efficiently. One of its key advancements is the ability to use tools and interact with software environments to complete multi-step processes. The model supports very large context windows, allowing it to analyze long documents and maintain context across extended conversations. GPT-5.4 also improves web research capabilities by searching and synthesizing information from multiple sources more effectively. Enhanced accuracy reduces hallucinations and helps produce more reliable responses for professional use. The model is available through ChatGPT, developer APIs, and coding environments such as Codex. By combining reasoning, tool usage, and large-scale context understanding, GPT-5.4 enables users to automate complex workflows and produce high-quality outputs.

MiMo-V2-Omni is a powerful multimodal AI model engineered to process and understand multiple types of data, including text, code, and structured inputs, within a unified system. It is designed to power agent-based workflows, enabling the execution of complex, multi-step tasks with improved accuracy and efficiency. The model combines advanced reasoning capabilities with strong tool integration, allowing it to interact with external systems and automate workflows effectively. It supports a wide range of applications, from software development and data analysis to enterprise automation and research tasks. With enhanced contextual understanding, it can maintain coherence across long interactions and complex scenarios. MiMo-V2-Omni is optimized for real-world performance, ensuring reliability in practical use cases rather than just benchmark results. Its architecture enables efficient handling of large-scale tasks while maintaining speed and responsiveness. The model also supports seamless integration into existing platforms and workflows. By combining multimodal understanding with agentic execution, it provides a flexible and scalable solution for modern AI applications. Overall, it delivers a balance of intelligence, versatility, and efficiency for diverse use cases.

Claude Mythos Preview is a next-generation language model designed with exceptional capabilities in cybersecurity analysis and exploit development. It has demonstrated the ability to autonomously identify zero-day vulnerabilities in major operating systems, web browsers, and widely used software. The model can go beyond detection by constructing functional exploits, including remote code execution and privilege escalation chains. It uses agentic workflows to explore codebases, test vulnerabilities, and validate findings without human intervention. Mythos Preview can also reverse engineer closed-source binaries, reconstructing logic and identifying potential weaknesses. Compared to earlier models, it shows a dramatic improvement in exploit success rates and complexity handling. The model is capable of chaining multiple vulnerabilities together to bypass modern security defenses. It can assist both defenders and attackers, depending on how it is used, highlighting the dual-use nature of advanced AI systems. These capabilities have led to initiatives focused on strengthening cybersecurity defenses using the model. Overall, Claude Mythos Preview represents a major advancement in AI-driven security research and automation.

Gemini 4 is Google’s upcoming next-generation AI model family and the future successor to its current Gemini 3.x lineup. Google has confirmed that it has started pre-training Gemini 4, describing it as its most ambitious pre-training run so far. The model has not been officially launched, so there are no public API endpoints, pricing details, model cards, benchmark results, or release dates available yet. Gemini 4 is expected to build on the direction of recent Gemini models, including stronger coding, reasoning, multimodal performance, computer use, and agentic workflow support. Google’s current Gemini releases already emphasize efficiency, lower latency, better tool use, and enterprise deployment, and Gemini 4 is likely to extend those priorities at a larger scale. The model may eventually support developers building AI agents, enterprise copilots, coding tools, research assistants, multimodal apps, and knowledge-work automation. It will likely play a role across Google AI Studio, the Gemini API, Gemini Enterprise, Google Cloud, and consumer-facing Gemini experiences once released. For now, Gemini 4 should be treated as a confirmed future model in training rather than an available product. By representing Google’s next major frontier model effort, Gemini 4 signals the company’s continued push toward more capable AI systems for developers, enterprises, and everyday users.

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.

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.

Baidu’s ERNIE 4.5 Turbo represents the next step in multimodal AI capabilities, combining advanced reasoning with the ability to process diverse forms of media like text, images, and audio. The model’s improved logical reasoning and memory retention ensure that businesses and developers can rely on more accurate outputs, whether for content generation, enterprise solutions, or educational tools. Despite its advanced features, ERNIE 4.5 Turbo is an affordable solution, priced at just a fraction of the competition. Baidu also plans to release this model as open-source in 2025, fostering greater accessibility for developers worldwide.

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.