24 stories tagged open-source-models, newest first.
54 points by Gcarter08 | 23 comments
Hugging Face model card for Hurston-1, an Apache-2.0 narrowband text-to-speech model fine-tuned from Maya1 for Southern American and African English accents over phone lines.
264 points by gmays | 74 comments
Blog post from Strands Agents announcing Strands Decider 2B, a small open-source decision model for fast experimentation and local development.
692 points by player85 | 189 comments
Blog post by the developer of Olly, an iMessage AI assistant, describing pitfalls and lessons learned running open models through OpenRouter's varied providers.
1918 points by mfiguiere | 888 comments
Business Insider exclusive reporting that Nvidia has been in talks to acquire AI model repository Hugging Face for more than $13 billion.
415 points by moonikakiss | 114 comments
Neon blog post describing how a 4B open-source model post-trained with Castform matches GPT-5.6 Sol on search-result retrieval accuracy at about 100x lower cost.
346 points by tosh | 169 comments
GitHub repository for Microsoft's VibeVoice, an open-source frontier voice AI project covering TTS and ASR models.
730 points by Poudlardo | 178 comments
Mistral AI announcement of Leanstral, an open-source code agent for Lean 4 aimed at trustworthy coding and formal proof engineering.
372 points by fugu2 | 197 comments
Blog post by Tim Plaisted explaining how to connect Claude Code to a local open-source LLM when the usage quota is exhausted.
939 points by pierre | 226 comments
GitHub repository for DeepSeek-OCR, DeepSeek's open-source "Contexts Optical Compression" OCR model, with code, paper PDF, and MIT license.
406 points by lastdong | 150 comments
Project page for Microsoft's VibeVoice, an open-source TTS model generating up to 90 minutes of multi-speaker conversational audio using LLM and diffusion.
897 points by divamgupta | 343 comments
GitHub repository for Kitten TTS, an open-source, CPU-only text-to-speech model under 25MB with 15M parameters and eight English voices.
630 points by pinter69 | 182 comments
GitHub repository for Chatterbox, Resemble AI's open-source state-of-the-art text-to-speech model released under the MIT license.
330 points by PaulineGar | 193 comments
Komiko web tool page for AniSora, Bilibili's open-source anime video generation model, letting users upload an image and prompt to generate animations.
362 points by csantini | 219 comments
Blog post introducing Audiblez, a tool that converts EPUB e-books into audiobooks using the Kokoro-82M open text-to-speech model.
573 points by pil0u | 71 comments
arXiv paper introducing OpenCoder, an open code LLM released with its training data, data pipeline, and protocols as an open cookbook for research.
627 points by CuriouslyC | 185 comments
Fal.ai blog post announcing availability of Flux, a 12B-parameter open-source text-to-image model, on the fal platform.
337 points by pentagrama | 137 comments
Ollama blog post from February 15, 2024 announcing a preview release of Ollama for Windows.
324 points by thecalebf | 157 comments
GitHub repository for Natural-SQL, a series of text-to-SQL language models including Natural-SQL-7B, with notebooks and prompt files.
500 points by artex_xh | 188 comments
Jina AI newsroom announcement of jina-embeddings-v2, an open-source text embedding model with 8K context length positioned as rivaling OpenAI's embeddings.
344 points by cgwu | 374 comments
Substack post by Vikram Sreekanti and Joseph E. Gonzalez arguing OpenAI's infrastructure scale makes its models too cheap for competitors to beat.
339 points by convexstrictly | 77 comments
A tweet by Thomas Wolf announcing that the Falcon 40B large language model is now licensed under Apache 2.0.
414 points by sadiq | 144 comments
GitHub repository for OpenLLaMA, an Apache-2.0 licensed open reproduction of Meta's LLaMA language model, with a 200B-token preview release.
1412 points by davidbarker | 474 comments
Stability AI blog post announcing the release of StableLM, the first models in its open-source language model suite.
805 points by tim_sw | 202 comments
Together blog post announcing RedPajama, a project to build open-source LLMs that starts by reproducing the LLaMA training dataset of over 1.2 trillion tokens.