DeepSeek has released models including DeepSeek-V4 and DeepSeek-R1; its official site states these models are open source, with model repositories and weight information on GitHub.
Gemma is a family of open models from Google DeepMind; according to Google, developers can use it to build apps on cloud servers, computers, phones and other environments.
GPT-6 Astra is a GPT-6-generation model released by OpenAI; according to OpenAI, it is available in some ChatGPT plans and the API as well as on Azure and Bedrock, and the line was later extended with versions such as GPT-6 Sol, Luna and GPT-6.1.
Tencent Hunyuan (Tencent HY) is a general-purpose and multimodal large model family fully self-developed by Tencent, covering modalities such as text and 3D, and providing enterprise-grade services for content production and business automation.
IBM Granite is a family of open models for enterprises; according to IBM, it includes language, vision, speech and guardrail models that can be customized and deployed.
Meta's Llama is a family of open models; the official documentation covers Llama 4, Llama 3 and Llama 2 and explains how to get the models and run them locally or in the cloud.
Meta's Muse model family includes Muse Spark, available through the Meta Model API, and Muse Glimmer, an open-weights model that runs locally; the official docs also list open-weights versions such as Muse Spark Open.
NVIDIA Nemotron is NVIDIA's family of multimodal foundation models that, according to NVIDIA, is aimed at tasks such as agentic reasoning, math and visual understanding.
Qwen (Tongyi Qianwen) is the large-model family from the Alibaba Cloud team; its official Hugging Face organization page says the team continues to release large language models, multimodal models and related projects.
Only necessary storage is on unless you agree. See the Privacy policy.
Necessary storage (sign-in session, your cookie choice, language preference, navigation expand state) is always on. Analytics and advertising are off by default and only turn on if you agree; this site currently loads no analytics or ad scripts. See the Privacy policy.