Top Results for Open Weight

Stable Diffusion 1.5 is an open-weight text-to-image artificial intelligence model released in October 2022 by Stability AI in collaboration with Runway and LMU Munich. It utilizes a latent diffusion architecture that generates detailed images by gradually denoising a mathematical representation of...

DeepSeek-V3 is a large language model released by the Chinese AI company DeepSeek in late 2024. It is built as a Mixture-of-Experts (MoE) model with a total of 671 billion parameters, of which only 37 billion are activated per token. The model was trained efficiently using a specialized architecture...

Llama 3.1 405B is a large language model released by Meta in 2024, serving as the flagship of the Llama 3.1 collection. It features 405 billion parameters and supports a context window of 128,000 tokens. As an open-weight model, it was made available for download, providing developers with a tool co...

Flux.1 Dev is a text-to-image diffusion model released in 2024 by Black Forest Labs, the company founded by former members of Stability AI. It is a guidance-distilled variant of the Flux.1 Pro model, released under a non-commercial license for research and development. The model uses a flow-matching...

Released in 2024 by the AI firm DeepSeek, DeepSeek Coder V2 is an open-weight, mixture-of-experts large language model designed for code generation and software engineering tasks. The model features a massive parameter count with an activated subset of experts for each query, enabling it to process...

Qwen2.5 is an open-weight family of large language models released in 2024 by Alibaba Cloud. The series includes base and instruction-tuned models ranging in size from 0.5 billion to 72 billion parameters. Notable for its strong multilingual capabilities and coding proficiency, the architecture is d...

QwQ-32B is a 32-billion-parameter large language model developed by Alibaba's Qwen team and released in late November 2024 under the Apache 2.0 license. It is designed as a reasoning-focused model that produces extended chain-of-thought outputs before delivering final answers, with reported strength...

Wan 2.1 is a text-to-video generation model developed by Alibaba and released as open-source in early 2025. The model utilizes a Diffusion Transformer (DiT) architecture to synthesize high-resolution video content directly from text prompts or reference images. It is available in multiple parameter...

Llama 3.3 is an instruction-tuned large language model developed by Meta and released in December 2024. Built with 70 billion parameters, it utilizes the same architecture as the larger Llama 3.1 405B model but achieves comparable performance through advanced post-training techniques. The model supp...

InternVL2 is an open-source vision-language foundation model developed by the Shanghai AI Laboratory, released in 2024. It is designed to process and reason across both visual and textual data, integrating a vision encoder with a large language model. The architecture is available in various paramet...

Gemma 2 27B is an open-weight large language model developed by Google DeepMind and released in June 2024. As the larger variant in the Gemma 2 family, it operates using 27 billion parameters and incorporates architectural optimizations such as knowledge distillation from larger proprietary models....

Mistral 7B is an open-weight large language model released in September 2023 by the French artificial intelligence company Mistral AI. The model features 7 billion parameters and utilizes architectural efficiencies like Grouped-Query Attention and Sliding Window Attention to accelerate processing an...

HunyuanVideo is an open-weight text-to-video generation model developed by Tencent and released in late 2024. Built on a diffusion transformer architecture, the model features 13 billion parameters, making it one of the largest publicly available video generation models of its time. It translates te...

Qwen2 is a family of large language models introduced by Alibaba Cloud's Qwen team in 2024. The series includes models at several parameter scales, uses a mixture of dense and mixture-of-experts architectures, and supports context processing and generation across numerous languages, including many b...

Gemma 2 is Google's second generation family of open-weight language models, introduced in 2024. It was released in 2-billion, 9-billion, and 27-billion parameter versions, allowing developers and researchers to choose among different memory and computing requirements for local or hosted deployment....

Tulu 3 is an open-weight instruction-tuned model series released by the Allen Institute for AI (AI2) in 2024. It is built on the Llama foundation model and incorporates post-training techniques including direct preference optimization (DPO) and reinforcement learning from verifiable rewards (RLVR) t...

DBRX is an open-weight mixture-of-experts (MoE) language model released by Databricks in March 2024. It has 132 billion total parameters with 36 billion active per token, utilizing 16 expert groups in a fine-grained routing architecture. Databricks trained the model on its own infrastructure and rep...

Stable Diffusion 3.5 is a text-to-image generation model released by Stability AI in 2024 as a refinement of Stable Diffusion 3. The release included multiple size variants, including Large, Large Turbo, and Medium, with publicly available model weights. It was designed to improve prompt adherence,...

LLaVA 1.6 is an open-weight, vision-language model developed by researchers from the University of Wisconsin–Madison and collaborating institutions. Released in early 2024, this iteration improves upon LLaVA 1.5 by supporting higher-resolution image inputs, which significantly enhances its optical c...

WizardLM 2 is a series of open-weight, instruction-tuned large language models developed by Microsoft and released in 2024. The series utilizes an evolved instruction-following approach to enhance the model's ability to handle complex queries and multi-turn conversations. Its 8x22B variant is freque...

OLMo 2 is a fully open-source large language model developed by the Allen Institute for AI (AI2) and released in 2024. As the second generation of the OLMo project, it provides the research community with open access to model weights, training datasets, intermediate checkpoints, and training code. T...

BLOOM is an open-access, multilingual large language model developed by the BigScience international research collaboration. Released in 2022, the model contains 176 billion parameters and was trained on text spanning 46 natural languages and 13 programming languages. Its primary purpose is to democ...

Llama 3.2 is a family of open-weight artificial intelligence models released by Meta in 2024. The release introduces lightweight text-only models with 1 billion and 3 billion parameters optimized for edge computing and mobile devices, alongside larger multimodal models that integrate computer vision...

Vicuna-13B is an open-weight large language model released in 2023 by LMSYS Org, a research organization comprising members from UC Berkeley, CMU, Stanford, and UCSD. The model was created by fine-tuning Meta's LLaMA-13B architecture on approximately 70,000 user-shared conversations sourced from Sha...

Aya 23 is an open-weight, generative artificial intelligence model developed by the Canadian startup Cohere and released in 2024. Available in 8-billion and 35-billion parameter configurations, the model natively supports 23 languages, including Arabic, Hindi, Spanish, and Chinese. It was released a...

LLaVA 1.5 is an open-weight multimodal large language model developed by researchers at the University of Wisconsin-Madison and released in 2023. The architecture connects a CLIP vision encoder to the Vicuna language model through an MLP projection layer, enabling the model to process and reason abo...

InternLM2.5 is an open-weight large language model series developed by Shanghai AI Laboratory and released in 2024 as part of their continued InternLM lineup. The series includes variants with different parameter sizes and a flagship model capable of processing up to 1 million tokens of context. Int...

Yi-34B is a 34-billion-parameter language model developed by 01.AI, an artificial intelligence company founded by Kai-Fu Lee, and released in 2023. The model is explicitly designed as bilingual, with strong capabilities in both English and Chinese across reasoning, mathematics, coding, and general k...

Mistral Nemo is a 12-billion-parameter open-weight large language model developed by Mistral AI in collaboration with NVIDIA and released in 2024. The model features a 128,000-token context window and was trained using a new tokenizer called Tekken that improves efficiency across multiple languages....