525 articles
OpenAI text-embedding-3
text-embedding-3 is a family of OpenAI text embedding models released in January 2024, in small and large versions that convert text into vectors for search, clustering, and retrieval-augmented generation.
OpenChat
OpenChat is a series of open-source large language models fine-tuned for conversation, led by Guan Wang. Its compact 7B models run on consumer GPUs and use the C-RLFT training method.
OpenELM
OpenELM is a family of four small open-weight language models released by Apple in April 2024, in 270 million to 3 billion parameter sizes, each with an instruction-tuned variant.
OpenFold
OpenFold is an open-source, trainable reproduction of DeepMind's AlphaFold2 protein-structure prediction model, built by a non-profit consortium and extended into OpenFold3 for cofolding proteins, nucleic acids, and small molecules.
OpenHands
OpenHands, formerly OpenDevin, is an open-source platform for AI coding agents, begun in March 2024 and developed by the company OpenHands, co-founded by Graham Neubig.
OpenHermes
OpenHermes is a series of open-source instruction-tuned large language models and an instruction dataset created by Teknium, best known for OpenHermes-2.5-Mistral-7B, a Mistral 7B fine-tune released in November 2023.
OpenLLaMA
OpenLLaMA is an open source, Apache 2.0 licensed reproduction of Meta AI's LLaMA language model, developed by OpenLM Research and trained on the open RedPajama dataset.
OpenMDW license
The OpenMDW license (Open Model, Data, and Weights) is a permissive, weights-first license for AI model distributions, published in 2025 by the Linux Foundation and used by NVIDIA for 19 models.
OpenRAIL licenses
OpenRAIL licenses, or Responsible AI Licenses, are AI model licenses that pair royalty-free open access to model weights with prohibited-use restrictions that bind every downstream derivative.
OpenVLA
OpenVLA is a 7-billion-parameter open-source vision-language-action model for robot manipulation, released in June 2024 and trained on 970,000 real-robot trajectories from the Open X-Embodiment dataset.
OpenVoice
OpenVoice is an open-source instant voice-cloning model family created by MyShell, first deployed in May 2023, that replicates a speaker's tone color from a short audio clip.
OPT (Meta model family)
OPT (Open Pre-trained Transformer) is a family of language models released by Meta AI in May 2022, with sizes from 125 million to 175 billion parameters, roughly matching GPT-3.
OPT-175B
OPT-175B is a 175-billion-parameter language model built by Meta AI like GPT-3 and released to researchers in May 2022 with training code and a complete training logbook.
Orca (language models)
Orca is a family of small instruction-tuned language models from Microsoft Research that learns reasoning from GPT-4's explanation traces; its successor, Orca 2, released in November 2023, picks its own reasoning strategy per task.
PaLI / PaLI-X
PaLI and PaLI-X are Google's multilingual vision-language models for image captioning and question answering in over 100 languages, with PaLI-X reaching 55 billion parameters in 2023.
PaliGemma
PaliGemma is an open vision-language model family from Google, first released in May 2024, combining the SigLIP image encoder with a Gemma decoder for fine-tuning on vision-language tasks.
PaLM (model family)
PaLM, or Pathways Language Model, was a family of large language models developed by Google, announced in 2022, that powered Bard and over 25 Google products.
Pangu (盘古) (model family)
Pangu (盘古) is a family of large language models developed by Huawei's Noah's Ark Lab, first released in April 2021 and focused on Chinese-language enterprise applications.
Parakeet (AI model)
Parakeet is a family of open automatic speech recognition models from NVIDIA's NeMo toolkit, whose English v2 model topped the Hugging Face Open ASR Leaderboard in 2025.
Parti
Parti (Pathways Autoregressive Text-to-Image) is a text-to-image model published by Google Research in June 2022 that generates images token by token like a language model, scaled to 20 billion parameters but never publicly released.
Phi (small language model family)
Phi is a family of small language models developed by Microsoft, running from Phi-1 through Phi-4, built on curated textbook-quality training data and used on phones, laptops, and Copilot+ PCs.
Phi-3
Phi-3 is a family of open small language models developed by Microsoft, first released in April 2024 with phi-3-mini, a 3.8 billion parameter model rivals said competed with GPT-3.5.
Phi-4
Phi-4 is a 14-billion-parameter large language model released by Microsoft in December 2024, trained mostly on synthetic data to match larger models on math and STEM reasoning.
Pika (AI model)
Pika is a family of text-to-video, image-to-video, and video-editing AI models developed by Pika Labs, founded in 2023, best known for viral one-click effects such as inflate, melt, and cake-ify.
Piper TTS
Piper TTS is a fast, open-source neural text-to-speech system that runs entirely on local CPUs, including the Raspberry Pi, offering over 100 voices in more than 30 languages.
PixArt
PixArt is an open-weight text-to-image model family built on a Diffusion Transformer rather than a UNet, first released in October 2023 and known for training at a fraction of the usual cost.
Pixtral
Pixtral is a family of open-weight vision-language models released by Mistral AI in 2024: Pixtral 12B under Apache 2.0 and the 124B Pixtral Large, both since discontinued.
PixVerse
PixVerse is a family of proprietary AI video-generation models developed by Aishi Technology, a Singapore- and Beijing-headquartered startup founded in 2023 by former ByteDance and Tencent senior researchers.
Playground v2.5
Playground v2.5 is a 3-billion-parameter open-weights text-to-image diffusion model released by Playground in February 2024, using the Stable Diffusion XL architecture trained from scratch for higher aesthetic quality.
Poolside (model family)
Poolside is a family of open-weight code foundation models from the company Poolside, founded in 2023, whose Laguna models are trained with reinforcement learning that rewards code that actually executes.