Models

Alibaba Releases Qwen3.8-Max and 27B Open Weights

AlphaSignal • 4 days ago

Prime Intellect Fable 5 Closes 82% of Human Research Gap

AlphaSignal • 3 days ago

Z.ai Launches GLM-5.3 With Surprising Cyber Capabilities

Unite.AI • 4 days ago

Meta Debuts Muse Glimmer 30B for Local AI Agents

InfoQ AI • 4 days ago

Cactus Compute releases Needle 2 tool-calling model

MarkTechPost • 4 days ago

GitHub Copilot Integrates xAI's Grok 4.6 Coding Model

Unite.AI • 4 days ago

ByteDance Launches Seedance 2.5 Video Model on Runway

AlphaSignal • 3 days ago

vLLM Integrates DeepSeek DSpark for Faster Inference

AlphaSignal • 4 days ago

Pika Labs Launches Pika Audio to Undercut Rivals

AlphaSignal • 4 days ago

Suno releases Studio 2.0 with MIDI and AI chatbot

The Verge AI • 4 days ago

Developers Adapt Qwen3-0.6B for Complex Tool Calling

MarkTechPost • 3 days ago

Google lets users disable visible watermarks on AI media

TechCrunch AI • 4 days ago

KDnuggets Outlines Best Tasks for Small Language Models

KDnuggets • 1 day ago

Simon Willison Builds CORS Chat Web UI for Local APIs

Simon Willison • 3 days ago

LTX-2.5 brings multimodal directing tools to AI video

The Neuron • 4 days ago

Build a Python Web Scraper Powered by GPT-5.4-Nano

KDnuggets • 4 days ago

Alibaba Releases Qwen3.8-Max and 27B Open Weights

Alibaba has released the open-source weights for Qwen3.8-Max and Qwen3.8-27B under the Apache 2.0 license, giving developers local access to its most...

AlphaSignal|4 days ago

Prime Intellect Fable 5 Closes 82% of Human Research Gap

Prime Intellect's Fable 5 model closed 81.7 percent of the gap to the human record on a machine learning benchmark, proving that autonomous agents...

AlphaSignal|3 days ago

Z.ai Launches GLM-5.3 With Surprising Cyber Capabilities

Z.ai has launched GLM-5.3, an upgraded model that demonstrates unexpectedly powerful multi-stage cyber exploitation and coding capabilities achieved...

Unite.AI|4 days ago

Meta Debuts Muse Glimmer 30B for Local AI Agents

Meta AI Research has open-sourced Muse Glimmer, a 30-billion-parameter model designed to run complex autonomous agents locally on consumer hardware...

InfoQ AI|4 days ago

Cactus Compute releases Needle 2 tool-calling model

Cactus Compute has launched Needle 2, an open 45-million-parameter tool-calling model designed to run locally on highly constrained hardware using...

MarkTechPost|4 days ago

GitHub Copilot Integrates xAI's Grok 4.6 Coding Model

GitHub has integrated xAI’s newly released Grok 4.6 model into its Copilot assistant, giving developers access to a highly competitive,...

Unite.AI|4 days ago

ByteDance Launches Seedance 2.5 Video Model on Runway

ByteDance has launched its Seedance 2.5 video generation model on Runway, allowing creators to produce longer, highly referenced 30-second clips to...

AlphaSignal|3 days ago

vLLM Integrates DeepSeek DSpark for Faster Inference

vLLM has integrated DeepSeek's DSpark framework to enable adaptive verification, allowing developers to optimize speculative decoding dynamically...

AlphaSignal|4 days ago

Pika Labs Launches Pika Audio to Undercut Rivals

Pika Labs has launched Pika Audio, a suite of four foundation models priced significantly lower than rivals like ElevenLabs to make multimedia AI...

AlphaSignal|4 days ago

Suno releases Studio 2.0 with MIDI and AI chatbot

Suno has launched Studio 2.0, introducing MIDI support and an interactive AI chatbot to transform its generative music platform into a fully-featured...

The Verge AI|4 days ago

Developers Adapt Qwen3-0.6B for Complex Tool Calling

A new supervised fine-tuning pipeline demonstrates how to adapt the Qwen3-0.6B model for complex tool-calling tasks using the XYZ-Aquila-SFT dataset.

MarkTechPost|3 days ago

Google lets users disable visible watermarks on AI media

Google is rolling out an option to remove visible watermarks from its AI-generated media, a move that gives creators more flexibility while...

TechCrunch AI|4 days ago

KDnuggets Outlines Best Tasks for Small Language Models

KDnuggets has outlined how developers can successfully deploy local small language models of eight billion parameters or fewer by targeting tasks...

KDnuggets|1 day ago

Simon Willison Builds CORS Chat Web UI for Local APIs

Independent developer Simon Willison has built CORS Chat, a browser-based web interface that lets developers interact directly with any...

Simon Willison|3 days ago

LTX-2.5 brings multimodal directing tools to AI video

The release of LTX-2.5, a 22-billion-parameter open world model, allows creators to direct AI video using a mix of text, audio, and keyframes rather...

The Neuron|4 days ago

Build a Python Web Scraper Powered by GPT-5.4-Nano

Developers can now build a streamlined Python web scraper that uses OpenAI's gpt-5.4-nano model to extract clean, query-specific markdown answers...

KDnuggets|4 days ago