Models

Reka Debuts Rho-1 to Merge Reasoning, Video, and Robotics

AlphaSignal • 1 day ago

Reka AI Debuts Rho-1 Omni-Model for Video and Robotics

The Decoder • 1 day ago

AutoTrust JEV-27B-VL Beats GPT-4o in Image Scoring

AlphaSignal • 1 day ago

Google DeepMind releases EmbeddingGemma 2 for phones

AlphaSignal • 14 hrs ago

Aleph Alpha Debuts Kolibri Open-Weight AI Model

The Decoder • 1 day ago

Ai2 Open-Sources Fast AstaBrief 8B Report Model

Hugging Face Blog • 4 days ago

Google's Nano Banana 2.1 Outperforms Pro on Image Edits

AlphaSignal • 14 hrs ago

Cloudflare launches Clef decision models for AI agents

The Decoder • 4 days ago

NASA and IBM release lunar AI foundation model

The Decoder • 2 days ago

Google Launches Compact EmbeddingGemma 2 Model

The Decoder • 10 hrs ago

China Telecom Debuts Xing4.0 Coding Agent Model

AlphaSignal • 2 days ago

Arena Post-Training Boosts FLUX.2-dev and Ideogram 4

AlphaSignal • 4 days ago

Cantina Releases apex-flash-1 for Security Research

MarkTechPost • 2 days ago

Mininglamp Releases Open-Source VisionHOPE Backbone

AlphaSignal • 2 days ago

Higgsfield Launches Ideogram 4.5 for Precise Image Edits

AlphaSignal • 4 days ago

Microsoft AI Releases Fast MAI Speech and Voice Models

The Decoder • 4 days ago

ByteShape Quantizes Qwen3.8-27B Down to 8.8 GB

AlphaSignal • 2 days ago

OpenBMB Releases 324M Draft Model for MiniCPM5-2B

AlphaSignal • 3 days ago

Suno Launches AI Tool to Blend Speech and Music

The Decoder • 4 days ago

Musubi Releases PolicyLM-1.7B for Real-Time Moderation

TechCrunch AI • 9 hrs ago

FlashML Runs MiniMax H3 Video AI on 8 GB GPUs

AlphaSignal • 3 days ago

Microsoft DeBERTa-v3-small Beats RoBERTa

AlphaSignal • 3 days ago

Humanizer 12B Rewrites AI Text to Bypass Detectors

AlphaSignal • 1 day ago

OpenAI and Rivals Release Four Frontier AI Models

MarkTechPost • 2 days ago

Red Hat Halves Nemotron 3.5 Lightning Size with FP8

AlphaSignal • 3 days ago

Mirror Particle builds AI world model of human behavior

TechCrunch AI • 13 hrs ago

Lipflow Translates Silent Mouthing Into Mac Text

AlphaSignal • 1 day ago

Convai Innovations Launches Laya Decision Engine

MarkTechPost • 5 hrs ago

Alibaba Scales Qwen AI Models to 2.4 Trillion Parameters

MarkTechPost • 2 days ago

VieNeu-TTS v3 Turbo Brings 48 kHz Vietnamese Speech to CPUs

AlphaSignal • 3 days ago

Simon Willison Builds Plugin for OpenAI Decisions API

Simon Willison • 7 hrs ago

llm-mistral 0.16 Adds Mistral Large 4 Support

Simon Willison • 8 hrs ago

Reka Debuts Rho-1 to Merge Reasoning, Video, and Robotics

Reka has unveiled Rho-1, a 19-billion-parameter model that processes text, video, and robot actions in one checkpoint, eliminating the need to chain...

AlphaSignal|1 day ago

Reka AI Debuts Rho-1 Omni-Model for Video and Robotics

Reka AI has released Rho-1, a 19-billion-parameter omni-model that unifies text, video, and robot control in a single network to simplify how...

The Decoder|1 day ago

AutoTrust JEV-27B-VL Beats GPT-4o in Image Scoring

AutoTrust has launched JEV-27B-VL, an open-source vision-language model that evaluates images and text in a single forward pass to deliver highly...

AlphaSignal|1 day ago

Google DeepMind releases EmbeddingGemma 2 for phones

Google DeepMind has launched EmbeddingGemma 2, a modular 740-million-parameter model that enables fast, private multimodal search across text, audio,...

AlphaSignal|14 hrs ago

Aleph Alpha Debuts Kolibri Open-Weight AI Model

German startup Aleph Alpha has launched Kolibri, a highly efficient open-weight model designed to comply with the EU AI Act and bolster European...

The Decoder|1 day ago

Ai2 Open-Sources Fast AstaBrief 8B Report Model

The Allen Institute for AI has open-sourced AstaBrief 8B, a specialized model that generates cited scientific reports 3.5 times faster than...

Hugging Face Blog|4 days ago

Google's Nano Banana 2.1 Outperforms Pro on Image Edits

Google has released Nano Banana 2.1, a fast and budget-friendly image model that beats its Pro counterpart on editing tasks, giving developers...

AlphaSignal|14 hrs ago

Cloudflare launches Clef decision models for AI agents

Cloudflare has released Clef and Clef-flash, two ultra-fast decision models designed to let AI agents autonomously categorize data and take actions...

The Decoder|4 days ago

NASA and IBM release lunar AI foundation model

NASA and IBM have released an open-source lunar foundation model trained on 17 years of satellite data, offering scientists a powerful new tool to...

The Decoder|2 days ago

Google Launches Compact EmbeddingGemma 2 Model

Google has launched EmbeddingGemma 2, a highly efficient open model that brings powerful multimodal embedding capabilities directly to local devices...

The Decoder|10 hrs ago

China Telecom Debuts Xing4.0 Coding Agent Model

China Telecom has released Xing4.0-29B-A4B, an open-source mixture-of-experts model that allows developers to run a powerful coding agent locally...

AlphaSignal|2 days ago

Arena Post-Training Boosts FLUX.2-dev and Ideogram 4

Arena has launched a text-to-image post-training method that uses gated rewards to prevent reward hacking, driving FLUX.2-dev and Ideogram 4 to the...

AlphaSignal|4 days ago

Cantina Releases apex-flash-1 for Security Research

Cantina Security has launched apex-flash-1, an open-weights model that identifies software vulnerabilities at a fraction of the cost of proprietary...

MarkTechPost|2 days ago

Mininglamp Releases Open-Source VisionHOPE Backbone

Mininglamp Technology and its partners have open-sourced VisionHOPE, a new vision backbone that uses adaptive memory to outperform traditional vision...

AlphaSignal|2 days ago

Higgsfield Launches Ideogram 4.5 for Precise Image Edits

Higgsfield has integrated Ideogram 4.5, a new image editing model designed to prevent pixel drift and color shifts during repeated, multi-turn...

AlphaSignal|4 days ago

Microsoft AI Releases Fast MAI Speech and Voice Models

Microsoft AI has launched three new models designed to power ultra-low-latency voice agents, offering rapid speech-to-text transcription and highly...

The Decoder|4 days ago

ByteShape Quantizes Qwen3.8-27B Down to 8.8 GB

ByteShape has released five highly compressed GGUF versions of the Qwen3.8-27B vision-language model, allowing developers to run the massive model...

AlphaSignal|2 days ago

OpenBMB Releases 324M Draft Model for MiniCPM5-2B

OpenBMB has launched MiniCPM5-2B-DSpark, a 324-million-parameter draft model designed to significantly accelerate inference speeds for its...

AlphaSignal|3 days ago

Suno Launches AI Tool to Blend Speech and Music

AI music platform Suno has launched a new feature called Speech that generates spoken-word audio layered with matching background music, expanding...

The Decoder|4 days ago

Musubi Releases PolicyLM-1.7B for Real-Time Moderation

AI startup Musubi has launched PolicyLM-1.7B, an open-weights decision model designed to moderate online content in real time without requiring...

TechCrunch AI|9 hrs ago

FlashML Runs MiniMax H3 Video AI on 8 GB GPUs

FlashML-org has launched FreeVideo, an open-source engine that lets creators run the MiniMax H3 video generation model locally on consumer-grade GPUs...

AlphaSignal|3 days ago

Microsoft DeBERTa-v3-small Beats RoBERTa

Microsoft Research's DeBERTa-v3-small is gaining massive developer traction by matching the performance of models twice its size using a highly...

AlphaSignal|3 days ago

Humanizer 12B Rewrites AI Text to Bypass Detectors

The new open-weights Humanizer 12B model rewrites AI-generated text to sound natural, allowing developers to run local workflows that successfully...

AlphaSignal|1 day ago

OpenAI and Rivals Release Four Frontier AI Models

Anthropic, OpenAI, and Google DeepMind launched four frontier-class AI models in September, forcing developers to weigh complex trade-offs in...

MarkTechPost|2 days ago

Red Hat Halves Nemotron 3.5 Lightning Size with FP8

Red Hat AI has released an FP8-quantized version of NVIDIA's Nemotron 3.5 Lightning 30B model, lowering the hardware requirements for deploying the...

AlphaSignal|3 days ago

Mirror Particle builds AI world model of human behavior

San Francisco startup Mirror Particle is developing a foundation model designed to simulate and predict human behavior, offering brands deeper...

TechCrunch AI|13 hrs ago

Lipflow Translates Silent Mouthing Into Mac Text

Lipflow, a new open-source macOS application, uses local AI to translate silent lip movements into text, offering a private and noise-free dictation...

AlphaSignal|1 day ago

Convai Innovations Launches Laya Decision Engine

Convai Innovations has released Laya, an open-source decision engine that offers developers a fast, calibrated alternative for zero-shot text...

MarkTechPost|5 hrs ago

Alibaba Scales Qwen AI Models to 2.4 Trillion Parameters

Alibaba Cloud has scaled its Qwen AI family from a modest 7-billion-parameter model to a massive 2.4-trillion-parameter release, reshaping the...

MarkTechPost|2 days ago

VieNeu-TTS v3 Turbo Brings 48 kHz Vietnamese Speech to CPUs

The newly released VieNeu-TTS v3 Turbo model delivers high-fidelity 48 kHz Vietnamese speech synthesis on standard CPUs, offering developers an...

AlphaSignal|3 days ago

Simon Willison Builds Plugin for OpenAI Decisions API

Developer Simon Willison has released an open-source plugin for OpenAI's new Decisions API, allowing practitioners to easily run structured binary,...

Simon Willison|7 hrs ago

llm-mistral 0.16 Adds Mistral Large 4 Support

Developer Simon Willison has released llm-mistral 0.16, updating the plugin to support Mistral's new reasoning models, including the newly released...

Simon Willison|8 hrs ago