Meta Releases Glimmer AI Model for Local Agent Use
Meta announced the open-source Glimmer model, a 30-billion parameter AI capable of running locally on consumer hardware, as part of its push toward personal superintelligence.
Model Release
428 articles
New AI models as they ship — large language models, multimodal systems, and open-weight releases. Capabilities, context windows, pricing, and availability, summarized so you can compare what is actually new.
LATEST NEWS
-
-
OpenAI Launches GPT-5.6-Cyber for Cybersecurity Defenders
OpenAI introduces GPT-5.6-Cyber, a cybersecurity-specific model, to help defenders counter AI-powered attacks. The model completes 95% of advanced cyber tasks, compared to 1.5% for its predecessor.
-
Hugging Face and EleutherAI Launch FineBooks OCR Leaderboard
Hugging Face and EleutherAI released an OCR leaderboard evaluating 14 open models on 2,165 historical book pages from the Biodiversity Heritage Library.
-
Anthropic Enables Auto Mode for Claude Code by Default
Anthropic is enabling auto mode for Claude Code by default starting August 14, with testing showing it caught 89% of harmful actions compared to 13.6% with manual review.
-
Google Deepmind's WeatherNext Cyclones Predicts Cyclone Tracks and Intensity Simultaneously
Google Deepmind's WeatherNext Cyclones (WN-C) forecasts tropical cyclones with greater accuracy than specialized models, improving track and intensity predictions by up to 28 percent.
-
Google's DiffusionGemma Shows Text Diffusion Without Full Retraining
Google DeepMind released DiffusionGemma, a text diffusion model built by retrofitting Gemma 4 with less than 10% of the original training token budget, achieving 1,500 tokens per second.
-
Backflip AI Releases CAD Model Generator for 3D Scans
Backflip AI's new tool transforms 3D scans into editable CAD models in minutes, addressing a critical gap in digital manufacturing.
-
Grok Imagine Image 2.0 Launches as Quality Mode
Grok Imagine Image 2.0 is now generally available as Quality Mode on grok.com/imagine and iOS/Android apps, ranking second in text-to-image and image editing benchmarks.
-
OpenAI Acknowledges Astra's Potential Cyber Capabilities
OpenAI says Astra, an upcoming model, may have critical cybersecurity capabilities, based on recent evaluations.
-
Lattice Retrieval Model Embeds Wikipedia in 7 Minutes
Erik Kaunismäki trained a static embedding model, lattice-retrieval, scoring 0.4749 NDCG on decontaminated BEIR after fine-tuning, outperforming sentence-transformers/static-retrieval-mrl-en-v1.
-
Bytedance Developing AI Model With Up to Ten Trillion Parameters
Bytedance is training an AI model with up to ten trillion parameters, three times the size of Moonshot's Kimi K3.
-
Hugging Face Releases Ideogram-4 Inpainting Model with Modular Diffusers
Hugging Face's Ideogram-4 inpainting model, released in August 2026, offers fine-grained control through JSON bounding boxes and differential diffusion, according to the blog post.