Skip to content

Search stories, models, companies…Subscribe

Topic · Concept

Open models

Open models are AI models whose weights are published for anyone to download and run, usually under a license that permits some reuse. The term covers a wide range, from fully open-source releases to "open-weight" models with restrictive licenses, and the difference matters for what you can legally do with them.

50 stories · updated Oct 29

Follow via RSS Get Open models alerts

Key facts

Checked Oct 8

Open Source AI Definition: Version 1.0, published October 28, 2024 (OSI)
Required components (OSAID): Data information, code, parameters
Four freedoms (OSAID): Use, study, modify, share, for any purpose

Start here

Open source · GuideHow to run an open model locally with Ollama (October 2026)Install Ollama, pull a model, chat with it, call it from code, and keep the server off the public internet. Every command comes from Ollama's official documentation.Oct 9·8 min read

Industry · CompareFrontier APIs vs open models: what they really costHosted open models are cheap, but the frontier labs' small models now list even lower. Self-hosting pays off only at very high, steady volume. Verified prices and the arithmetic.Oct 9·7 min read

Open source · NewsLiquid AI LFM2.5-24B-A2B: a 24B open-weight MoE with 2B active parameters, its license and what it needs to runLiquid AI released LFM2.5-24B-A2B on October 13, 2026: a 24-billion-parameter mixture-of-experts reasoning model with about 2.45B active parameters and a 131K context. The weights are free to download, but Liquid's license excludes commercial use by companies with $10 million or more in annual revenue.Oct 14·5 min read

Latest on Open models

Open source · NewsNemotron 3.5 Transcribe: Nvidia's open speech recognition model covers 40 languages under OpenMDW, with about 4B active parametersNemotron 3.5 Transcribe, released by Nvidia on Hugging Face on October 28, 2026, is an open-weight speech recognition model for 40 languages under the OpenMDW-1.1 license. It activates about 4B parameters per token, runs on a single Nvidia GPU and handles recorded audio, not live streams. Nvidia's word error rates are vendor-reported.Oct 29·6 min read

Research · AnalysisQwenputer: a developer compiled a RISC-V CPU into Qwen3.5-4B's weights, with no training, and says it runs 18 million instructions without an errorQwenputer is Qwen3.5-4B with a 32-bit RISC-V CPU compiled directly into 11 extra layers, using the model's Gated DeltaNet memory as RAM. Developer Divyansh Agrawal says no weights were trained, the CPU passes the official RISC-V tests, and Qwen can call it as a tool for exact arithmetic. What was built, how, what 'deterministic' means and the limits.Oct 29·7 min read

Open source · ExplainerWhat is an abliterated model? How refusal removal works, why abliterated DeepSeek is now sold by API, and the safety questionsWhat is an abliterated model? It's an open-weight model whose weights have been edited to strip out its learned refusals, based on 2024 research showing refusal sits along a single direction inside the model. In October 2026, an API router began selling an abliterated DeepSeek V4.1 Flash from five providers. What the technique is, what the licenses and terms allow, and what safety researchers say.Oct 29·8 min read

Products · ExplainerOffline AI assistant: how Survival LLM pairs a local model with a 930,000-article emergency library, and how people run AI with no internetAn offline AI assistant runs the model and its reference library on your own computer, so it keeps working when the internet doesn't. Survival LLM, which drew wide attention on October 27, 2026, bundles one of four models with about 930,000 Wikipedia articles and medical and repair references. What it needs, how Kiwix, Ollama, llama.cpp and LM Studio work offline, and where offline AI falls short.Oct 29·9 min read

Open source · NewsLiquid AI LFM3: six 2B-class open-weight models for text, vision, audio, omni and decisions, free only below $10 million in revenueLiquid AI LFM3, released October 27, 2026, puts text, vision, audio, omni and two 'decision' models on one 1.7B backbone. What 'every modality' means, the sizes, the LFM Open License's $10 million revenue limit, and where the models run.Oct 29·6 min read

Industry · AnalysisWhy are AI model prices dropping? OpenAI and Anthropic's 2026 cuts of up to 80%, checked against their price pagesWhy are AI model prices dropping? Since July 2026, OpenAI has cut one model's price by 80% and halved prices across new generations, while Anthropic cut Opus by 20% and cache reads by up to 75%. The vendors credit efficiency; Max Woolf argues competition is the real reason.Oct 27·7 min read

Open source · ExplainerOpen-weight vs open-source AI models: what the difference is, which models qualify, and why IBM's CEO says open weights now win on costOpen-weight vs open-source models: an open-weight model gives you the trained parameters; an open-source AI model, under the OSI's definition, also gives you the training code and enough data information to rebuild it. Most popular 'open' models are only open-weight. Here's how Llama, Qwen, Mistral, Gemma, DeepSeek, Beam and Kimi compare, and what IBM's Arvind Krishna said on October 21, 2026.Oct 27·8 min read

Open source · NewsHurston-1 TTS: an Apache 2.0 voice model for Southern American and African-accented English, built for phone callsHurston-1 is a 3-billion-parameter text-to-speech model from Atlanta's Destined AI, released on Hugging Face under Apache 2.0 and shared on Hacker News on October 25, 2026. It speaks Southern American and several African English accents, with an 8 kHz phone mode. What it was trained on, what its makers report, and its stated limits.Oct 27·5 min read

Hardware · NewsHow Cerebras says it serves trillion-parameter models: 59 wafers for Kimi K3, pipelined layer by layerHow Cerebras serves trillion-parameter models, per its October 22, 2026 engineering post: Kimi K3's layers are split across 59 WSE-3T wafers with 44 GB of on-chip memory each, run as a pipeline, with a target of 1,000 to 2,000+ tokens per second. All figures are vendor-reported.Oct 26·4 min read

Hardware · NewsHugging Face's Microduck robot reaches 20,000 preorders: price, specs and what it doesHugging Face's Microduck robot passed 20,000 preorders after Jensen Huang's GTC Berlin demo, CEO Clem Delangue said, according to TechCrunch. The $399 biped from Pollen Robotics ships from before Christmas 2026 with Apache 2.0 software but closed hardware designs.Oct 26·5 min read

Nothing in this format on this page — try another filter or the next page.