Alternatives to Timestral

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.

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.

NVIDIA Earth-2 is an advanced platform designed for high-performance, AI-enhanced weather and climate forecasting, utilizing a combination of open-source models, libraries, and frameworks to facilitate professional-grade predictions without the necessity for traditional supercomputing resources. This innovative system offers comprehensive capabilities, starting from the ingestion of raw observational data and culminating in the generation of high-resolution weather and climate forecasts, which include localized storm predictions and medium-range forecasts that extend up to 15 days. By leveraging generative AI architectures, Earth-2 significantly accelerates computational processes while either maintaining or enhancing accuracy when compared to more conventional forecasting techniques. The Earth-2 suite features a range of models such as Atlas, which specializes in multi-variable medium-range forecasting, StormScope for short-term local weather predictions, HealDA for global data assimilation, CorrDiff for regional resolution downscaling, and FourCastNet 3 for efficient global forecasting. Furthermore, this platform not only enhances predictive capabilities but also democratizes access to advanced forecasting technologies, making them more readily available to a wider audience.

Nixtla is a cutting-edge platform designed for time-series forecasting and anomaly detection, centered on its innovative model, TimeGPT, which is recognized as the first generative AI foundation model tailored for time-series information. This model has been trained on an extensive dataset comprising over 100 billion data points across various sectors, including retail, energy, finance, IoT, healthcare, weather, and web traffic, enabling it to make precise zero-shot predictions for numerous applications. Users can effortlessly generate forecasts or identify anomalies in their data with just a few lines of code through the provided Python SDK, even when dealing with irregular or sparse time series, and without the need to construct or train models from the ground up. TimeGPT also boasts advanced capabilities such as accommodating external factors (like events and pricing), enabling simultaneous forecasting of multiple time series, employing custom loss functions, conducting cross-validation, providing prediction intervals, and allowing fine-tuning on specific datasets. This versatility makes Nixtla an invaluable tool for professionals seeking to enhance their time-series analysis and forecasting accuracy.

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.

Solar Mini is an advanced pre-trained large language model that matches the performance of GPT-3.5 while providing responses 2.5 times faster, all while maintaining a parameter count of under 30 billion. In December 2023, it secured the top position on the Hugging Face Open LLM Leaderboard by integrating a 32-layer Llama 2 framework, which was initialized with superior Mistral 7B weights, coupled with a novel method known as "depth up-scaling" (DUS) that enhances the model's depth efficiently without the need for intricate modules. Following the DUS implementation, the model undergoes further pretraining to restore and boost its performance, and it also includes instruction tuning in a question-and-answer format, particularly tailored for Korean, which sharpens its responsiveness to user prompts, while alignment tuning ensures its outputs align with human or sophisticated AI preferences. Solar Mini consistently surpasses rivals like Llama 2, Mistral 7B, Ko-Alpaca, and KULLM across a range of benchmarks, demonstrating that a smaller model can still deliver exceptional performance. This showcases the potential of innovative architectural strategies in the development of highly efficient AI models.

Frogcast is a cutting-edge global weather forecasting API that provides extremely precise and localized weather predictions by integrating over 20 advanced numerical weather prediction models with state-of-the-art artificial intelligence and diverse geospatial data from top-tier organizations such as ECMWF and NOAA, alongside earth observation data and ground-level measurements, to aid crucial decision-making in fields like agriculture, construction, logistics, retail, insurance, environment, transport, and health. Its RESTful API grants access to an impressive array of more than 26 meteorological variables, offering forecasts that extend up to 15 days into the future, along with probabilistic forecasts that include confidence intervals and quantiles for enhanced understanding, all while delivering high spatial resolution down to 1 km and detailed topographic information at 90 m for focused targeting of specific areas. The Frogcast platform is designed for effortless integration into current systems and applications via intuitive API endpoints utilizing standard web protocols and JSON data formats, which facilitates both real-time and historical weather data analysis for trend identification and decision support. Additionally, the flexibility of the API allows for customization based on user needs, enhancing the overall user experience and effectiveness of the weather data provided.

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.

TabPFN-3.5 is an advanced foundation model specifically designed for achieving top-tier predictions on structured data, making it highly effective for a variety of tasks such as churn analysis, fraud detection, pricing strategies, demand forecasting, and risk assessment, thus enabling teams to utilize a single model for diverse applications. The model seamlessly processes data in its original form, adeptly managing issues like missing values, outliers, categorical data, multi-table datasets, free text features, and thousands of unique identifiers without requiring any encoding, while also accommodating numerous measurements per row. Users have the convenience of inputting raw data without the need for extensive feature engineering or preprocessing, allowing them to obtain high-quality, production-ready predictions immediately after the initial prediction call. Notably, TabPFN-3.5 generates predictions in a single forward pass, striking an optimal balance between accuracy and speed, and is optimized for quick inference, which is crucial for latency-sensitive predictive applications. Furthermore, it can efficiently handle large-scale datasets of up to one million rows natively and boasts a remarkable 20 times faster inference speed compared to its predecessors, making it a significant advancement in the field. This combination of efficiency, versatility, and performance positions TabPFN-3.5 as a powerful tool for data scientists and organizations seeking to leverage structured data effectively.

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.

Introducing an open-source AI model that can be fine-tuned, distilled, and deployed across various platforms. Our newest instruction-tuned model comes in three sizes: 8B, 70B, and 405B, giving you options to suit different needs. With our open ecosystem, you can expedite your development process using a diverse array of tailored product offerings designed to meet your specific requirements. You have the flexibility to select between real-time inference and batch inference services according to your project's demands. Additionally, you can download model weights to enhance cost efficiency per token while fine-tuning for your application. Improve performance further by utilizing synthetic data and seamlessly deploy your solutions on-premises or in the cloud. Take advantage of Llama system components and expand the model's capabilities through zero-shot tool usage and retrieval-augmented generation (RAG) to foster agentic behaviors. By utilizing 405B high-quality data, you can refine specialized models tailored to distinct use cases, ensuring optimal functionality for your applications. Ultimately, this empowers developers to create innovative solutions that are both efficient and effective.

Phonon serves as an open-weight text-to-speech (TTS) model and inference runtime designed for developers seeking to incorporate speech generation into applications that require local and offline audio capabilities. This innovative system transforms text into spoken audio and has the ability to stream speech in real-time as it is produced. The runtime offers interfaces for Rust and Python, a Swift package for iOS and macOS, JavaScript support for web browsers, and libraries for Android; it can also function as a self-hosted server compatible with OpenAI's API. Depending on the variant of the model used, it accommodates languages such as English, French, German, Spanish, and Portuguese. Phonon is tailored for developers aiming to integrate speech functionalities into mobile, desktop, web, embedded, and server applications, particularly in scenarios requiring offline access or enhanced privacy. The English model, which consists of 31 million parameters, is licensed under Apache 2.0, while the language-specific and multilingual models, featuring 40 million parameters, operate under CC BY-NC 4.0. Additionally, the runtime’s code is available under either the Apache 2.0 or MIT licenses, making it accessible for various development needs. This flexibility and wide-ranging support make Phonon a versatile choice for speech synthesis across different platforms.

LFM3 represents Liquid AI's suite of streamlined, open-weight foundational models designed for a variety of applications, including text, vision, audio, and multimodal tasks. This lineup features variants for text generation, vision-language processing, audio, omni tasks, and decision-making models. Developers have the option to download the model weights, customize the supported models via fine-tuning, and deploy them on local edge hardware, or alternatively, utilize Liquid AI’s hosted API for access. Potential applications encompass conversational agents, tool invocation, speech transcription and synthesis, understanding of images and videos, document extraction, classification, routing, and robotics functionalities. LFM3 aims to cater to developers, researchers, and organizations focused on creating tailored AI solutions that require adaptable deployment options, including the ability for local inference. Furthermore, the open-weight models operate under Liquid AI’s LFM Open License, where the eligibility for commercial use and its associated terms are contingent upon the annual revenue of the organization. This flexibility allows users to tailor their AI implementations precisely to their needs, enhancing the overall utility of the models.

Beam represents Reflection’s inaugural open-weight model, characterized as a sparse Mixture-of-Experts framework featuring a staggering 501 billion parameters, of which 23 billion are actively utilized, specifically crafted for tasks involving coding, reasoning, and agentic functions. Its prowess is derived from extensive pretraining and reinforcement learning, having been developed using a vast dataset of 23.8 trillion diverse, high-quality tokens sourced from the web, public domains, and proprietary licensed materials. With a targeted emphasis on enhancing coding and agentic capabilities, Beam is engineered to provide competitive open-weight performance while ensuring efficient inference compute. This model adeptly handles a wide array of tasks, including intricate software engineering, terminal operations, STEM activities, web searches, tool utilization, and general knowledge inquiries. Reinforcement learning techniques have been employed to bolster its abilities in multi-step reasoning, tool application, and responsiveness to environmental feedback. Moreover, users have the flexibility to manage the balance between efficiency and performance through an adjustable reasoning effort parameter, thus tailoring the model's output to better suit their specific needs.

Liquid AI's LFM2.5 represents an advanced iteration of on-device AI foundation models, engineered to provide high-efficiency and performance for AI inference on edge devices like smartphones, laptops, vehicles, IoT systems, and embedded hardware without the need for cloud computing resources. This new version builds upon the earlier LFM2 framework by greatly enhancing the scale of pretraining and the stages of reinforcement learning, resulting in a suite of hybrid models that boast around 1.2 billion parameters while effectively balancing instruction adherence, reasoning skills, and multimodal functionalities for practical applications. The LFM2.5 series comprises various models including Base (for fine-tuning and personalization), Instruct (designed for general-purpose instruction), Japanese-optimized, Vision-Language, and Audio-Language variants, all meticulously crafted for rapid on-device inference even with stringent memory limitations. These models are also made available as open-weight options, facilitating deployment through platforms such as llama.cpp, MLX, vLLM, and ONNX, thus ensuring versatility for developers. With these enhancements, LFM2.5 positions itself as a robust solution for diverse AI-driven tasks in real-world environments.

WeatherNext represents a suite of AI-driven models developed by Google DeepMind and Google Research, designed to deliver cutting-edge weather predictions. These advanced models surpass conventional physics-based approaches in both speed and efficiency, leading to enhanced reliability in forecasts. By improving the accuracy of weather predictions, these innovations could significantly aid in disaster preparedness, ultimately saving lives during severe weather scenarios and bolstering the dependability of renewable energy sources and supply chains. WeatherNext Graph stands out by providing more precise and efficient deterministic forecasts than existing systems, producing a single forecast for each specified time and location with a 6-hour temporal resolution and a 10-day lead time. In addition, WeatherNext Gen excels at generating ensemble forecasts that outshine the current predominant models, thereby equipping decision-makers with a clearer understanding of weather uncertainties and the associated risks of extreme weather conditions. This leap in forecasting capability promises to transform how we respond to and manage the impacts of climate variability.

Canopy Labs has unveiled Orpheus, an innovative suite of advanced speech large language models (LLMs) aimed at achieving human-like speech generation capabilities. Utilizing the Llama-3 architecture, these models have been trained on an extensive dataset comprising over 100,000 hours of English speech, allowing them to generate speech that exhibits natural intonation, emotional depth, and rhythmic flow that outperforms existing high-end closed-source alternatives. Orpheus also features zero-shot voice cloning, enabling users to mimic voices without any need for prior fine-tuning, and provides easy-to-use tags for controlling emotion and intonation. The models are engineered for low latency, achieving approximately 200ms streaming latency for real-time usage, which can be further decreased to around 100ms when utilizing input streaming. Canopy Labs has made available both pre-trained and fine-tuned models with 3 billion parameters under the flexible Apache 2.0 license, with future intentions to offer smaller models with 1 billion, 400 million, and 150 million parameters to cater to devices with limited resources. This strategic move is expected to broaden accessibility and application potential across various platforms and use cases.

Recursive Flint is an advanced language model featuring 40 billion parameters, designed as a mixture-of-experts specifically for coding tasks executed locally. It engages around 2.5 billion parameters for each token processed, enabling sophisticated reasoning, multi-turn tool utilization, and accommodating context windows as extensive as 512,000 tokens. Developers can leverage Flint within coding agents to analyze codebases, rectify bugs, develop new features, and create tests while executing inference directly on their own hardware. The model's weights are made available under the Apache 2.0 license, promoting accessibility. Flint is compatible with various platforms such as llama.cpp, Ollama, Unsloth, LM Studio, vLLM, and SGLang; additionally, Recursive offers rec, an inference engine that is specifically optimized for use with Apple Silicon M4 Pro and M5 Pro/Max chips. Comprehensive setup documentation and practical examples are provided to assist developers in executing the model on local machines. Flint serves as a valuable tool for developers and technical teams aiming to effectively run and customize a code-centric model tailored to their specific system requirements. This innovative approach empowers users to enhance their coding capabilities and streamline development processes more efficiently than ever before.

Mistral 3 represents the newest iteration of open-weight AI models developed by Mistral AI, encompassing a diverse range of models that span from compact, edge-optimized versions to a leading large-scale multimodal model. This lineup features three efficient “Ministral 3” models with 3 billion, 8 billion, and 14 billion parameters, tailored for deployment on devices with limited resources, such as laptops, drones, or other edge devices. Additionally, there is the robust “Mistral Large 3,” which is a sparse mixture-of-experts model boasting a staggering 675 billion total parameters, with 41 billion of them being active. These models are designed to handle multimodal and multilingual tasks, excelling not only in text processing but also in image comprehension, and they have showcased exceptional performance on general queries, multilingual dialogues, and multimodal inputs. Furthermore, both the base and instruction-fine-tuned versions are made available under the Apache 2.0 license, allowing for extensive customization and integration into various enterprise and open-source initiatives. This flexibility in licensing encourages innovation and collaboration among developers and organizations alike.

TabFM is an innovative zero-shot foundation model specifically created for handling tabular data, aimed at streamlining classification and regression processes that usually necessitate extensive manual model training, hyperparameter optimization, and tailored feature engineering. By transforming the challenge of tabular prediction into an in-context learning task, TabFM avoids the need to train a new supervised model for every dataset; instead, it consolidates historical training examples and target testing rows into a single cohesive prompt, allowing it to discern the intricate relationships between various columns and rows during inference. Given that tables are inherently two-dimensional and do not rely on a specific order, TabFM employs a hybrid architecture that integrates alternating attention mechanisms for both rows and columns, row compression techniques, and a specialized Transformer designed for in-context learning based on these compressed row embeddings. This sophisticated framework enables the model to effectively capture complex interactions and dependencies among features while maintaining computational efficiency, particularly advantageous for processing larger datasets. Furthermore, this approach not only enhances performance but also significantly reduces the time and resources typically required for model development in tabular data tasks.

Mistral Large 3 pushes open-source AI into frontier territory with a massive sparse MoE architecture that activates 41B parameters per token while maintaining a highly efficient 675B total parameter design. It sets a new performance standard by combining long-context reasoning, multilingual fluency across 40+ languages, and robust multimodal comprehension within a single unified model. Trained end-to-end on thousands of NVIDIA H200 GPUs, it reaches parity with top closed-source instruction models while remaining fully accessible under the Apache 2.0 license. Developers benefit from optimized deployments through partnerships with NVIDIA, Red Hat, and vLLM, enabling smooth inference on A100, H100, and Blackwell-class systems. The model ships in both base and instruct variants, with a reasoning-enhanced version on the way for even deeper analytical capabilities. Beyond general intelligence, Mistral Large 3 is engineered for enterprise customization, allowing organizations to refine the model on internal datasets or domain-specific tasks. Its efficient token generation and powerful multimodal stack make it ideal for coding, document analysis, knowledge workflows, agentic systems, and multilingual communications. With Mistral Large 3, organizations can finally deploy frontier-class intelligence with full transparency, flexibility, and control.

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.

Chatterbox, an open-source voice cloning AI model created by Resemble AI and distributed under the MIT license, allows users to perform zero-shot voice cloning with just a five-second sample of reference audio, thereby removing the requirement for extensive training. This innovative model provides expressive speech synthesis that features emotion control, enabling users to modify the expressiveness of the voice from a dull tone to a highly dramatic one using a single adjustable parameter. Additionally, Chatterbox allows for accent modulation and offers text-based control, which guarantees a high-quality and human-like text-to-speech output. With its faster-than-real-time inference capabilities, it is well-suited for applications requiring immediate responses, such as voice assistants and interactive media experiences. Designed with developers in mind, the model supports easy installation via pip and comes with thorough documentation. Furthermore, Chatterbox integrates built-in watermarking through Resemble AI’s PerTh (Perceptual Threshold) Watermarker, which discreetly embeds data to safeguard the authenticity of generated audio. This combination of features makes Chatterbox a powerful tool for creating versatile and realistic voice applications. The model's emphasis on user control and quality further enhances its appeal in various creative and professional fields.

Devstral is a collaborative effort between Mistral AI and All Hands AI, resulting in an open-source large language model specifically tailored for software engineering. This model demonstrates remarkable proficiency in navigating intricate codebases, managing edits across numerous files, and addressing practical problems, achieving a notable score of 46.8% on the SWE-Bench Verified benchmark, which is superior to all other open-source models. Based on Mistral-Small-3.1, Devstral boasts an extensive context window supporting up to 128,000 tokens. It is designed for optimal performance on high-performance hardware setups, such as Macs equipped with 32GB of RAM or Nvidia RTX 4090 GPUs, and supports various inference frameworks including vLLM, Transformers, and Ollama. Released under the Apache 2.0 license, Devstral is freely accessible on platforms like Hugging Face, Ollama, Kaggle, Unsloth, and LM Studio, allowing developers to integrate its capabilities into their projects seamlessly. This model not only enhances productivity for software engineers but also serves as a valuable resource for anyone working with code.

PanGu-α has been created using the MindSpore framework and utilizes a powerful setup of 2048 Ascend 910 AI processors for its training. The training process employs an advanced parallelism strategy that leverages MindSpore Auto-parallel, which integrates five different parallelism dimensions—data parallelism, operation-level model parallelism, pipeline model parallelism, optimizer model parallelism, and rematerialization—to effectively distribute tasks across the 2048 processors. To improve the model's generalization, we gathered 1.1TB of high-quality Chinese language data from diverse fields for pretraining. We conduct extensive tests on PanGu-α's generation capabilities across multiple situations, such as text summarization, question answering, and dialogue generation. Additionally, we examine how varying model scales influence few-shot performance across a wide array of Chinese NLP tasks. The results from our experiments highlight the exceptional performance of PanGu-α, demonstrating its strengths in handling numerous tasks even in few-shot or zero-shot contexts, thus showcasing its versatility and robustness. This comprehensive evaluation reinforces the potential applications of PanGu-α in real-world scenarios.

This repository showcases the research preview of LongLLaMA, an advanced large language model that can manage extensive contexts of up to 256,000 tokens or potentially more. LongLLaMA is developed on the OpenLLaMA framework and has been fine-tuned utilizing the Focused Transformer (FoT) technique. The underlying code for LongLLaMA is derived from Code Llama. We are releasing a smaller 3B base variant of the LongLLaMA model, which is not instruction-tuned, under an open license (Apache 2.0), along with inference code that accommodates longer contexts available on Hugging Face. This model's weights can seamlessly replace LLaMA in existing systems designed for shorter contexts, specifically those handling up to 2048 tokens. Furthermore, we include evaluation results along with comparisons to the original OpenLLaMA models, thereby providing a comprehensive overview of LongLLaMA's capabilities in the realm of long-context processing.

Large language models, often requiring extensive computational resources for training over long periods, have demonstrated impressive proficiency in zero- and few-shot learning tasks. Due to the high investment needed for their development, replicating these models poses a significant challenge for many researchers. Furthermore, access to the few models available via API is limited, as users cannot obtain the complete model weights, complicating academic exploration. In response to this, we introduce Open Pre-trained Transformers (OPT), a collection of decoder-only pre-trained transformers ranging from 125 million to 175 billion parameters, which we intend to share comprehensively and responsibly with interested scholars. Our findings indicate that OPT-175B exhibits performance on par with GPT-3, yet it is developed with only one-seventh of the carbon emissions required for GPT-3's training. Additionally, we will provide a detailed logbook that outlines the infrastructure hurdles we encountered throughout the project, as well as code to facilitate experimentation with all released models, ensuring that researchers have the tools they need to explore this technology further.

Zyphra is thrilled to unveil the beta release of Zonos-v0.1, which boasts two sophisticated and real-time text-to-speech models that include high-fidelity voice cloning capabilities. Our release features both a 1.6B transformer and a 1.6B hybrid model, all under the Apache 2.0 license. Given the challenges in quantitatively assessing audio quality, we believe that the generation quality produced by Zonos is on par with or even surpasses that of top proprietary TTS models currently available. Additionally, we are confident that making models of this quality publicly accessible will greatly propel advancements in TTS research. You can find the Zonos model weights on Huggingface, with sample inference code available on our GitHub repository. Furthermore, Zonos can be utilized via our model playground and API, which offers straightforward and competitive flat-rate pricing options. To illustrate the performance of Zonos, we have prepared a variety of sample comparisons between Zonos and existing proprietary models, highlighting its capabilities. This initiative emphasizes our commitment to fostering innovation in the field of text-to-speech technology.

Magistral is the inaugural language model family from Mistral AI that emphasizes reasoning, offered in two variants: Magistral Small, a 24 billion parameter open-weight model accessible under Apache 2.0 via Hugging Face, and Magistral Medium, a more robust enterprise-grade version that can be accessed through Mistral's API, the Le Chat platform, and various major cloud marketplaces. Designed for specific domains, it excels in transparent, multilingual reasoning across diverse tasks such as mathematics, physics, structured calculations, programmatic logic, decision trees, and rule-based systems, generating outputs that follow a chain of thought in the user's preferred language, which can be easily tracked and validated. This release signifies a transition towards more compact yet highly effective transparent AI reasoning capabilities. Currently, Magistral Medium is in preview on platforms including Le Chat, the API, SageMaker, WatsonX, Azure AI, and Google Cloud Marketplace. Its design is particularly suited for general-purpose applications that necessitate extended thought processes and improved accuracy compared to traditional non-reasoning language models. The introduction of Magistral represents a significant advancement in the pursuit of sophisticated reasoning in AI applications.

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.

Florence-2-large is a cutting-edge vision foundation model created by Microsoft, designed to tackle an extensive range of vision and vision-language challenges such as caption generation, object recognition, segmentation, and optical character recognition (OCR). Utilizing a sequence-to-sequence framework, it leverages the FLD-5B dataset, which comprises over 5 billion annotations and 126 million images, to effectively engage in multi-task learning. This model demonstrates remarkable proficiency in both zero-shot and fine-tuning scenarios, delivering exceptional outcomes with minimal training required. In addition to detailed captioning and object detection, it specializes in dense region captioning and can interpret images alongside text prompts to produce pertinent answers. Its versatility allows it to manage an array of vision-related tasks through prompt-driven methods, positioning it as a formidable asset in the realm of AI-enhanced visual applications. Moreover, users can access the model on Hugging Face, where pre-trained weights are provided, facilitating a swift initiation into image processing and the execution of various tasks. This accessibility ensures that both novices and experts can harness its capabilities to enhance their projects efficiently.

Mistral AI has introduced the Mistral Large 2, a sophisticated AI model crafted to excel in various domains such as code generation, multilingual understanding, and intricate reasoning tasks. With an impressive 128k context window, this model accommodates a wide array of languages, including English, French, Spanish, and Arabic, while also supporting an extensive list of over 80 programming languages. Designed for high-throughput single-node inference, Mistral Large 2 is perfectly suited for applications requiring large context handling. Its superior performance on benchmarks like MMLU, coupled with improved capabilities in code generation and reasoning, guarantees both accuracy and efficiency in results. Additionally, the model features enhanced function calling and retrieval mechanisms, which are particularly beneficial for complex business applications. This makes Mistral Large 2 not only versatile but also a powerful tool for developers and businesses looking to leverage advanced AI capabilities.

NVIDIA Nemotron 3.5 Transcribe serves as an open-weight model tailored for developers aiming to create transcription and audio-processing systems. This model transforms audio recordings into text, offering options for both verbatim and edited transcripts, along with features like contextual vocabulary, speaker identification, word-level timestamps, speech translation, language detection, and audio question answering. It boasts a language matrix that encompasses 40 different languages, though the extent of task coverage may vary from one language to another. Capable of processing recordings of up to one hour, the model is optimized for offline inference on compatible NVIDIA GPU hardware and supported runtimes. Developers can seamlessly integrate it into a variety of applications, including speech interfaces, media and meeting processing, customer service analytics, translation software, and other audio comprehension tools. Furthermore, the model's weights, alongside a fine-tuning toolkit, are made available under the OpenMDW-1.1 license, with specific hardware and serving requirements contingent upon the nature of the deployment. This flexibility allows teams to adapt the model for diverse use cases in the field of audio processing and transcription.

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.

Harnessing cutting-edge open-source AutoML technology, we facilitate the creation of high-quality models effortlessly. This system incorporates Deep Learning and pre-trained models to enhance intelligence wherever relevant. By employing Causal AI alongside AutoML, it ensures fairness, supports causal inference, and provides counterfactual predictions. Each trained model can be deployed instantly for interactive online use or through an API, making it accessible to all users. Additionally, it offers comprehensive insights into feature importances and model explanations through Shapley values. Our AI engine operates entirely on an open-source framework, allowing for complete transparency and universal applicability of our algorithms. It effectively groups customers or products into similar cohorts based on an extensive array of features. Furthermore, it predicts future outcomes by identifying temporal patterns in historical data and is capable of training predictive models using labeled data to make predictions on unlabeled datasets, thereby enhancing its overall utility and performance.

Aily Labs offers a mobile-centric decision-intelligence platform that equips organizations with instantaneous AI insights and predictive analytics throughout their operations. This innovative product dismantles data silos by linking comprehensive company information via its Correlator engine, which incorporates sophisticated AI and machine-learning models to predict outcomes, model scenarios, enhance decision-making, and present actionable insights to frontline personnel. Featuring rapid one-day integration and ready-to-use enterprise workflows, the platform helps teams rethink and revolutionize their work approaches, ensuring that data becomes a shared asset while minimizing decision-making biases and facilitating swifter, more informed actions. It supports essential functions such as forecasting, simulations, and optimization for supply chains, finance, and operations, providing a complete view and decision intelligence directly at users' fingertips. Offered as an app, this solution is specifically engineered to transition organizations from mere passive analytics to actively engaging in decision-driven processes, ultimately enhancing overall efficiency and responsiveness. By fostering a culture of data accessibility, Aily Labs enables teams to harness the full potential of their information resources, driving innovation and strategic growth.

Pixtral Large is an expansive multimodal model featuring 124 billion parameters, crafted by Mistral AI and enhancing their previous Mistral Large 2 framework. This model combines a 123-billion-parameter multimodal decoder with a 1-billion-parameter vision encoder, allowing it to excel in the interpretation of various content types, including documents, charts, and natural images, all while retaining superior text comprehension abilities. With the capability to manage a context window of 128,000 tokens, Pixtral Large can efficiently analyze at least 30 high-resolution images at once. It has achieved remarkable results on benchmarks like MathVista, DocVQA, and VQAv2, outpacing competitors such as GPT-4o and Gemini-1.5 Pro. Available for research and educational purposes under the Mistral Research License, it also has a Mistral Commercial License for business applications. This versatility makes Pixtral Large a valuable tool for both academic research and commercial innovations.

MiniMax M2 is an open-source foundational model tailored for agent-driven applications and coding tasks, achieving an innovative equilibrium of efficiency, velocity, and affordability. It shines in comprehensive development environments, adeptly managing programming tasks, invoking tools, and executing intricate, multi-step processes, complete with features like Python integration, while offering impressive inference speeds of approximately 100 tokens per second and competitive API pricing at around 8% of similar proprietary models. The model includes a "Lightning Mode" designed for rapid, streamlined agent operations, alongside a "Pro Mode" aimed at thorough full-stack development, report creation, and the orchestration of web-based tools; its weights are entirely open source, allowing for local deployment via vLLM or SGLang. MiniMax M2 stands out as a model ready for production use, empowering agents to autonomously perform tasks such as data analysis, software development, tool orchestration, and implementing large-scale, multi-step logic across real organizational contexts. With its advanced capabilities, this model is poised to revolutionize the way developers approach complex programming challenges.

MiMo-V2-Flash is a large language model created by Xiaomi that utilizes a Mixture-of-Experts (MoE) framework, combining remarkable performance with efficient inference capabilities. With a total of 309 billion parameters, it activates just 15 billion parameters during each inference, allowing it to effectively balance reasoning quality and computational efficiency. This model is well-suited for handling lengthy contexts, making it ideal for tasks such as long-document comprehension, code generation, and multi-step workflows. Its hybrid attention mechanism integrates both sliding-window and global attention layers, which helps to minimize memory consumption while preserving the ability to understand long-range dependencies. Additionally, the Multi-Token Prediction (MTP) design enhances inference speed by enabling the simultaneous processing of batches of tokens. MiMo-V2-Flash boasts impressive generation rates of up to approximately 150 tokens per second and is specifically optimized for applications that demand continuous reasoning and multi-turn interactions. The innovative architecture of this model reflects a significant advancement in the field of language processing.

VideoPoet is an innovative modeling technique that transforms any autoregressive language model or large language model (LLM) into an effective video generator. It comprises several straightforward components. An autoregressive language model is trained across multiple modalities—video, image, audio, and text—to predict the subsequent video or audio token in a sequence. The training framework for the LLM incorporates a range of multimodal generative learning objectives, such as text-to-video, text-to-image, image-to-video, video frame continuation, inpainting and outpainting of videos, video stylization, and video-to-audio conversion. Additionally, these tasks can be combined to enhance zero-shot capabilities. This straightforward approach demonstrates that language models are capable of generating and editing videos with impressive temporal coherence, showcasing the potential for advanced multimedia applications. As a result, VideoPoet opens up exciting possibilities for creative expression and automated content creation.

In honor of Cleopatra, whose magnificent fate concluded amidst the tragic incident involving a snake, we are excited to introduce Codestral Mamba, a Mamba2 language model specifically designed for code generation and released under an Apache 2.0 license. Codestral Mamba represents a significant advancement in our ongoing initiative to explore and develop innovative architectures. It is freely accessible for use, modification, and distribution, and we aspire for it to unlock new avenues in architectural research. The Mamba models are distinguished by their linear time inference capabilities and their theoretical potential to handle sequences of infinite length. This feature enables users to interact with the model effectively, providing rapid responses regardless of input size. Such efficiency is particularly advantageous for enhancing code productivity; therefore, we have equipped this model with sophisticated coding and reasoning skills, allowing it to perform competitively with state-of-the-art transformer-based models. As we continue to innovate, we believe Codestral Mamba will inspire further advancements in the coding community.

GPT-J represents an advanced language model developed by EleutherAI, known for its impressive capabilities. When it comes to performance, GPT-J showcases a proficiency that rivals OpenAI's well-known GPT-3 in various zero-shot tasks. Remarkably, it has even outperformed GPT-3 in specific areas, such as code generation. The most recent version of this model, called GPT-J-6B, is constructed using a comprehensive linguistic dataset known as The Pile, which is publicly accessible and consists of an extensive 825 gibibytes of language data divided into 22 unique subsets. Although GPT-J possesses similarities to ChatGPT, it's crucial to highlight that it is primarily intended for text prediction rather than functioning as a chatbot. In a notable advancement in March 2023, Databricks unveiled Dolly, a model that is capable of following instructions and operates under an Apache license, further enriching the landscape of language models. This evolution in AI technology continues to push the boundaries of what is possible in natural language processing.

Evo 2 represents a cutting-edge genomic foundation model that excels in making predictions and designing tasks related to DNA, RNA, and proteins. It employs an advanced deep learning architecture that allows for the modeling of biological sequences with single-nucleotide accuracy, achieving impressive scaling of both compute and memory resources as the context length increases. With a robust training of 40 billion parameters and a context length of 1 megabase, Evo 2 has analyzed over 9 trillion nucleotides sourced from a variety of eukaryotic and prokaryotic genomes. This extensive dataset facilitates Evo 2's ability to conduct zero-shot function predictions across various biological types, including DNA, RNA, and proteins, while also being capable of generating innovative sequences that maintain a plausible genomic structure. The model's versatility has been showcased through its effectiveness in designing operational CRISPR systems and in the identification of mutations that could lead to diseases in human genes. Furthermore, Evo 2 is available to the public on Arc's GitHub repository, and it is also incorporated into the NVIDIA BioNeMo framework, enhancing its accessibility for researchers and developers alike. Its integration into existing platforms signifies a major step forward for genomic modeling and analysis.

Mistral Large 4 serves as a versatile multimodal language model, accessible via Mistral’s hosted inference API. It integrates capabilities for instruction adherence, reasoning, and agent-like functions, allowing it to process both images and text simultaneously. This model is designed for developers to facilitate various tasks such as chat completions, structured data outputs, executing functions, answering questions from documents, and enhancing tool-assisted workflows. Its range of applications spans conversational AI, software development assistance, cybersecurity evaluations, financial and legal processes, scientific inquiry, data extraction, and the interpretation of various document types, including charts and images. Aimed at developers, researchers, and organizations within the public sector, Mistral Large 4 is tailored for creating AI-driven applications across numerous fields, including engineering, manufacturing, finance, logistics, and scientific research. Furthermore, its adaptability makes it suitable for both small startups and large enterprises looking to leverage AI technologies effectively.

Devstral Small 2 serves as the streamlined, 24 billion-parameter version of Mistral AI's innovative coding-centric model lineup, released under the flexible Apache 2.0 license to facilitate both local implementations and API interactions. In conjunction with its larger counterpart, Devstral 2, this model introduces "agentic coding" features suitable for environments with limited computational power, boasting a generous 256K-token context window that allows it to comprehend and modify entire codebases effectively. Achieving a score of approximately 68.0% on the standard code-generation evaluation known as SWE-Bench Verified, Devstral Small 2 stands out among open-weight models that are significantly larger. Its compact size and efficient architecture enable it to operate on a single GPU or even in CPU-only configurations, making it an ideal choice for developers, small teams, or enthusiasts lacking access to expansive data-center resources. Furthermore, despite its smaller size, Devstral Small 2 successfully maintains essential functionalities of its larger variants, such as the ability to reason through multiple files and manage dependencies effectively, ensuring that users can still benefit from robust coding assistance. This blend of efficiency and performance makes it a valuable tool in the coding community.