All AI news
Browse, filter, and search every article in the archive. The homepage shows the last 24 hours; everything older lives here.
Sina's open model VibeThinker-3B aims to show reasoning compresses well but factual knowledge doesn't
Sina just launched VibeThinker-3B, an open model that demonstrates how reasoning compresses effectively while factual knowledge does not. This could impact how developers approach model training and knowledge retention in AI systems.
Asian AI startups launch Mythos-like models as Anthropic’s export ban drags on
Asian AI startups are launching models similar to Anthropic's Mythos while facing export bans. This move indicates a growing competitive landscape in AI development despite regulatory hurdles.
Anthropic gets US approval to bring back Claude Mythos 5
Anthropic just received US approval to reintroduce Claude Mythos 5. This means users can expect enhanced capabilities and features from the AI model soon.
OpenAI's new flagship model GPT-5.6 Sol cheats on software tests more than any model before it
OpenAI just released GPT-5.6 Sol, which outperforms previous models in software tests. This means users can expect more accurate and efficient coding assistance from the model.
ByteDance's "iLLaDA" is a diffusion language model that keeps up with Qwen2.5
ByteDance just launched 'iLLaDA', a new diffusion language model that competes with Qwen2.5. This model aims to enhance language processing capabilities in various applications, pushing the boundaries of AI-generated content.
[AINews] OpenAI GPT-5.6 Sol / Terra / Luna — restricted to trusted partners
OpenAI just restricted access to GPT-5.6 Sol / Terra / Luna to trusted partners. This means only select users can leverage the latest model capabilities, impacting how widely the technology can be adopted.
Anthropic’s Mythos 5 is back
Anthropic just reintroduced Mythos 5 with improved capabilities and safety features. This update aims to enhance user experience by making interactions more reliable and aligned with user intentions.
OpenAI's GPT-5.6 Sol launches to rival Claude Mythos under government access rules it calls unsustainable
OpenAI just launched GPT-5.6 Sol to compete with Claude Mythos while navigating new government access rules. This move aims to enhance user access and capabilities in a rapidly evolving regulatory landscape.
OpenAI Has New AI Models. Here’s Why You Can’t Use Them
OpenAI just released new AI models but restricted access for users. This means developers can't use them yet, limiting innovation and integration opportunities.
OpenAI unveils GPT-5.6 amid US AI regulatory drama
OpenAI just unveiled GPT-5.6 while navigating US AI regulatory challenges. This update aims to enhance performance and compliance, making it more viable for various applications amid changing regulations.
A man of many words
A new AI model from MIT generates text that mimics human-like conversation. This advancement could enhance chatbots and virtual assistants, making interactions feel more natural.
[AINews] GLM > GPT? GLM-5.2 passes vibe check; Z.ai forecasts Open Fable by December
GLM-5.2 just passed evaluations, showing strong performance against GPT models. This means users might see a viable alternative for their AI needs soon.
GLM-5.2 is probably the most powerful text-only open weights LLM
GLM-5.2 just launched as a powerful text-only open weights LLM. It offers enhanced performance for developers looking to integrate advanced language capabilities into their applications.
[AINews] GLM-5.2: the top Frontend Coding model in the world, IndexShare for Speculative Decoding
GLM-5.2 just launched as the leading frontend coding model. This upgrade enhances coding efficiency and accuracy for developers working on complex projects.
Introducing Gemma 4 models on Amazon Bedrock
Amazon just launched Gemma 4 models on Bedrock. Users can now access advanced AI capabilities for building and deploying applications more efficiently.
New AI model called "Count Anything" does exactly what it says, and that's harder than it sounds
A new AI model called 'Count Anything' can accurately count various objects in images. This capability enhances image analysis and could improve applications in inventory management and visual data processing.
Google Research's Gemini-SQL2 tops text-to-SQL benchmarks by a wide margin
Google Research just launched Gemini-SQL2, which outperforms competitors in text-to-SQL benchmarks. This means developers can expect more accurate and efficient SQL query generation from natural language inputs.
Microsoft's SkillOpt boosts GPT-5.5 by using nothing but a trained Markdown file
Microsoft just boosted GPT-5.5 with SkillOpt, leveraging a trained Markdown file for enhanced performance. This means users can expect more efficient and accurate outputs from the model.
Claude Fable 5 outpaces GPT-5.5 by 13 points on FrontierMath's toughest problems
Claude Fable 5 just outperformed GPT-5.5 by 13 points on FrontierMath's toughest problems. This shows Claude's growing strength in tackling complex mathematical challenges, which could enhance its utility in educational and professional settings.
Google DeepMind releases DiffusionGemma, a model that runs local AI 4x faster
Google DeepMind just released DiffusionGemma, a model that runs local AI four times faster. This boost in speed enhances performance for users running AI tasks on their own devices.