Local Model Watch is a news site covering open-weight generative models you can run on your own hardware \u2014 text, image, video and audio \u2014 with a focus on reporting news quickly, in both Japanese and English.<\/p>\n
Local Model Watch uses AI agents to streamline the collection, organization and writing of news about local AI.<\/p>\n
The operator (an individual) designs and maintains the automation itself, the sources we follow, the publication criteria, how articles are structured, and how corrections are handled.<\/p>\n
We track inference engines, quantization and fine-tuning tools, as well as new releases and updates for models that generate text, images, video or audio. Models that don’t generate anything \u2014 embeddings, classifiers and the like \u2014 are out of scope.<\/p>\n
We follow announcements from well-known projects and organizations (major inference engines and model publishers), as well as topics that clearly show a spike in attention on developer communities such as Hacker News and Reddit.<\/p>\n
Wherever possible, articles cite official announcements and primary sources. We never publish an article that has no source links.<\/strong> How articles are produced, how figures are computed and how corrections are handled is described in our Editorial Policy<\/a>.<\/p>\n To choose which model our writing agents themselves use, we refer to model performance data published by Artificial Analysis<\/a>. This data is only used behind the scenes for that selection and is never included in article content.<\/p>\n If an article turns out to be wrong, we do not delete it. We add a correction notice at the top of the article and update it so that what was wrong and what was corrected both stay visible. See our Privacy Policy<\/a> for more.<\/p>\nWhat you will only find here<\/h2>\n
\n
\n requirements and GPU tiers computed by this site from the actual size of the distributed files \u2014 not figures quoted from the model card.<\/li>\n
\n have covered, grouped by the smallest VRAM tier it is estimated to run in.<\/li>\n
\n original article are appended to that article.<\/li>\n
\n articles, tracked variants and hardware requirements.<\/li>\n
\n vLLM and the other projects we watch, the full release history alongside our articles \u2014 including the smaller releases that do not get an article of their own.<\/li>\n
\n covered models with a similar parameter count for comparison.<\/li>\n
\n quantization and model formats<\/a> \u2014 the same definitions our writing pipeline works from.<\/li>\n
\n articles.<\/li>\n<\/ul>\nServices we use<\/h2>\n
When we get something wrong<\/h2>\n
Get in touch<\/h2>\n