Every AI News story tagged with both Open-Source AI and AI Research — the two topics side by side, updated as new articles publish.
15 articles
PrismML has released Bonsai 27B, a 27-billion-parameter open-weight model compressed to under 4GB, small enough to run natively on a phone while keeping most of its reasoning ability.
Read more →Mira Murati's startup released Inkling, a 975-billion-parameter open-weight model built for developers to fine-tune and adapt, rather than to top leaderboards.
Read more →
DeepSeek is the Chinese AI lab whose cheaply trained, open-weight models triggered Nvidia's biggest one-day stock drop and reshaped how the industry talks about AI costs.
Read more →GLM is the open-weight AI model family from Chinese lab Zhipu AI (Z.ai) — free to download and self-host, and competitive with top US models at a fraction of the API cost.
Read more →Z.ai has released GLM-5.2, a 744-billion-parameter open-weight model that matches GPT-5.5 on key benchmarks and trails Claude Opus 4.8 narrowly — trained entirely on Huawei silicon with no Nvidia hardware.
Read more →Mistral's new open-source model for Lean 4 fully saturates the miniF2F theorem benchmark, solves 87% of graduate-level math tests, and found five previously unknown bugs in production code — at roughly $4 per proof versus $300 for competing systems.
Read more →Chinese technology company Meituan has released LongCat-2.0, a 1.6-trillion-parameter coding model trained entirely on domestic Chinese chips, under the MIT License — the first frontier-class open-source model built without Western semiconductors.
Read more →Open-weight AI models release their trained parameters publicly so anyone can download, run, or fine-tune them — unlike closed models such as GPT-4 or Gemini, which are accessible only through paid APIs. Here is what that difference means in practice.
Read more →OpenAI on June 22 expanded its Daybreak security program with GPT-5.5-Cyber, a model designed to detect and patch code vulnerabilities, and launched Patch the Planet to harden over 30 major open-source projects.
Read more →Open-source AI models release their trained weights publicly, letting anyone run, study, or customize them without paying API fees. This guide explains what that means in practice and how to get started.
Read more →Liquid AI released LFM2.5-230M, a 230-million-parameter open-weight model that runs AI agent tasks at 213 tokens per second on a phone CPU and 42 tokens per second on a Raspberry Pi — without a cloud connection.
Read more →Chinese lab Zhipu AI has open-sourced GLM-5.2, a 753-billion-parameter model that outperforms GPT-5.5 on software engineering benchmarks while costing roughly one-sixth as much.
Read more →Alibaba's Qwen team has released Qwen-AgentWorld, a family of open-source language world models that predict how digital environments respond to AI agent actions, with its 397B variant outscoring GPT-5.4 on the new AgentWorldBench benchmark.
Read more →The Google DeepMind spinoff signed a three-year, $6.3 billion computing contract with SpaceX's Colossus data center, giving it the GPU scale needed to challenge closed AI giants with open-weight models.
Read more →OpenAI and Trail of Bits launched Patch the Planet, using the GPT-5.5-Cyber model to find and fix vulnerabilities in widely used open-source software, with 30+ projects on board.
Read more →