All AI news
Browse, filter, and search every article in the archive. The homepage shows the last 24 hours; everything older lives here.
Google’s new anything-to-anything AI model is wild
Google just launched a new anything-to-anything AI model that can generate diverse outputs from various inputs. This model expands creative possibilities for developers and users, enabling more versatile applications across different domains.
Towards Speed-of-Light Text Generation with Nemotron-Labs Diffusion Language Models
Hugging Face just introduced Nemotron-Labs, a new diffusion language model aimed at achieving speed-of-light text generation. This model promises to enhance the efficiency and responsiveness of AI-generated content, making it more practical for real-time applications.
Perplexity Is Open-Sourcing Bumblebee
Perplexity is open-sourcing Bumblebee, their new AI model designed for enhanced conversational capabilities. This move allows developers to customize and integrate Bumblebee into their applications, boosting accessibility and innovation in AI interactions.
llm-gemini 0.32
Gemini just released version 0.32 with improved performance and new features. Users can expect faster processing and enhanced capabilities for handling complex tasks.
Gemini 3.5 Flash: more expensive, but Google plan to use it for everything
Google is rolling out Gemini 3.5, which comes with a higher price tag. This move signals their intent to integrate it across all their services, enhancing AI capabilities throughout their ecosystem.
llm-gemini 0.32a0
Gemini just released version 0.32a0, improving its performance and capabilities. Users can expect enhanced efficiency and more accurate outputs in their applications.
[AINews] Codex Rises, Claude Meters Programmatic Usage
OpenAI is enhancing Codex to improve its programmatic usage and reduce irrelevant outputs. This update aims to make Codex more effective for developers in real-world applications.
[AINews] The End of Finetuning
Latent Space just announced a new approach that eliminates the need for finetuning in AI models. This change simplifies the training process and could lead to faster deployment of AI solutions.
llm 0.32a2
Simon Willison just released LLM 0.32a2, an update to his language model. This version includes improved performance and new features for developers working with LLMs.
[AINews] Thinking Machines' Native Interaction Models - TML-Interaction-Small 276B-A12B - advances SOTA Realtime Voice and kills standard VAD
Thinking Machines just launched TML-Interaction-Small 276B-A12B, advancing state-of-the-art real-time voice interaction. This model replaces standard Voice Activity Detection (VAD) for smoother and more responsive voice applications.
OpenAI just released its answer to Claude Mythos
OpenAI just launched its new model, GPT-5, to compete with Anthropic's Claude Mythos. This release enhances capabilities for developers and businesses looking for advanced AI solutions.
GPT-5.5 costs 49 to 92 percent more than its predecessor, depending on the input length
OpenAI is launching GPT-5.5, which costs 49 to 92 percent more than GPT-5 based on input length. This price increase reflects the enhanced capabilities and resources required for the new model.
CyberSecQwen-4B: Why Defensive Cyber Needs Small, Specialized, Locally-Runnable Models
Hugging Face just launched CyberSecQwen-4B, a specialized model for defensive cybersecurity. This model allows organizations to run AI locally, enhancing security by minimizing data exposure to external servers.
Chrome's 4GB AI model isn't new, but you're not wrong for being confused
Google is clarifying that its 4GB AI model isn't a new release despite some confusion. This means users can expect consistent performance without the need to adapt to a new model.
EMO: Pretraining mixture of experts for emergent modularity
Hugging Face just introduced EMO, a pretraining mixture of experts model designed for emergent modularity. This approach enhances model efficiency and adaptability, allowing for more specialized tasks without requiring extensive retraining.
Running Codex safely at OpenAI
OpenAI is enhancing Codex to prevent it from generating irrelevant content. This update aims to make Codex more effective for real-world applications.
v2.1.105
Claude just released version 2.1.105 with various bug fixes and performance improvements. Users can expect a smoother experience and more reliable outputs from the AI.
v2.1.101
Claude just released version 2.1.101, which includes various performance improvements and bug fixes. Users can expect a smoother experience and enhanced reliability in their interactions.
v2.1.98
Claude just released version 2.1.98 with various bug fixes and performance improvements. Users can expect a smoother experience and enhanced reliability in their interactions.
llm-gemini 0.31
Gemini just released version 0.31, enhancing its capabilities with improved performance and new features. Users can expect a more robust experience with better handling of complex tasks and queries.