Models

Meta Releases Muse Wonder to Reclaim AI Frontier

WIRED AI • 1 day ago

Mistral's Le Chonk Beats GPT-6-Astra in Image Test

AlphaSignal • 1 day ago

Microsoft Launches Decision-1 Scoring Model

AlphaSignal • 4 days ago

OpenAI Debuts GPT-6 Featuring Interactive UI

Mindstream AI • 4 days ago

Cloudflare Launches Clef-omni Decision Model

MarkTechPost • 1 day ago

Google Tests Gemini 4 Carbon Variant to Boost Coding

The Decoder • 3 days ago

Speridlabs Releases Iris-3B Pixel-Space Generator

AlphaSignal • 1 day ago

Alibaba releases 8-step Qwen-Image-2.1-Turbo model

MarkTechPost • 4 days ago

Conway Research Releases Saluki 27B for Local AI Agents

MarkTechPost • 4 days ago

Z.ai Opens Access to 320B GLM-Legal AI Model

AlphaSignal • 1 day ago

Jina Releases Multimodal jina-embeddings-v5-omni-small

AlphaSignal • 4 days ago

TII Releases Falcon-ASR Arabic Speech Model

MarkTechPost • 1 day ago

OpenAI Launches Composer Predictions in Codex

AlphaSignal • 4 days ago

vLLM Semantic Router Releases Decision 3.0

MarkTechPost • 2 days ago

Nace AI Open-Sources Drex 1.5 Decision Model

MarkTechPost • 3 days ago

Reflection AI and Mistral Release New Open-Weight Models

AI Business • 4 days ago

OrcaRouter Releases OrcaCyber Zero 1.5 Security Model

MarkTechPost • 2 days ago

PatientDesk AI Releases 30MB Antalia-2 Mini Speech Model

AlphaSignal • 2 days ago

AutoTrust AI Launches 45ms GEV-26B Decision Model

AlphaSignal • 2 days ago

Google DeepMind Unifies Seven Tracking Models in TapNet

AlphaSignal • 3 days ago

ACE-Step 1.5 Brings Local Music Generation to C++

AlphaSignal • 3 days ago

UkisAI Shrinks Swift 1.5 27B Model to 8.42GB

AlphaSignal • 1 day ago

SC117 Strips Refusals From Qwen3.8-Flash-Next

AlphaSignal • 2 days ago

Alibaba's 4-Bit Qwen2.5-Coder Hits 775K Downloads

AlphaSignal • 4 days ago

Bartowski Shrinks Qwen2.5-32B Model Down to 9GB

AlphaSignal • 1 day ago

DigUp Uses EmbeddingGemma 2 for Local Mac Search

AlphaSignal • 3 days ago

Hugging Face Adds RepVGG-A0 Model to timm

AlphaSignal • 3 days ago

Google Launches Playground AI to Generate Web Video Games

WIRED AI • 2 days ago

Umi-OCR Hits 47k Stars for Local Desktop Text Extraction

AlphaSignal • 3 days ago

Simon Willison Updates ttok Tool for GPT-6 Tokenization

Simon Willison • 4 days ago

Meta Releases Muse Wonder to Reclaim AI Frontier

Meta has unveiled its Muse Wonder AI models, positioning the tech giant back at the forefront of artificial intelligence after heavy investments in...

WIRED AI|1 day ago

Mistral's Le Chonk Beats GPT-6-Astra in Image Test

Mistral has launched Mistral Large 4, nicknamed Le Chonk, showcasing top visual grounding capabilities that allow AI systems to pinpoint exact image...

AlphaSignal|1 day ago

Microsoft Launches Decision-1 Scoring Model

Microsoft has launched Microsoft-Decision-1, a specialized decision-scoring model designed to drastically speed up AI classification, routing, and...

AlphaSignal|4 days ago

OpenAI Debuts GPT-6 Featuring Interactive UI

OpenAI has launched GPT-6 for ChatGPT, introducing interactive user interfaces and streaming responses to streamline workflows for its 1.2 billion...

Mindstream AI|4 days ago

Cloudflare Launches Clef-omni Decision Model

Cloudflare has launched Clef-omni, an open-weight 30B multimodal decision model that evaluates audio, video, text, and images in a single call to...

MarkTechPost|1 day ago

Google Tests Gemini 4 Carbon Variant to Boost Coding

Google is internally evaluating new Gemini 4 model variants like Carbon, which reportedly delivers coding capabilities competitive with Anthropic's...

The Decoder|3 days ago

Speridlabs Releases Iris-3B Pixel-Space Generator

Speridlabs has released Iris-3B, an open-source text-to-image model that generates megapixel images directly in pixel space without relying on a...

AlphaSignal|1 day ago

Alibaba releases 8-step Qwen-Image-2.1-Turbo model

Alibaba’s Qwen team has launched Qwen-Image-2.1-Turbo, an 8-step image model that drastically cuts generation steps while retaining 2K resolution and...

MarkTechPost|4 days ago

Conway Research Releases Saluki 27B for Local AI Agents

Conway Research has released Saluki 27B, a compressed open-source model that outperforms its 54 GB base model at local tool calling while running in...

MarkTechPost|4 days ago

Z.ai Opens Access to 320B GLM-Legal AI Model

Z.ai has opened gated access to GLM-Legal, a 320-billion-parameter domain-specific model designed to streamline contract reviews, compliance checks,...

AlphaSignal|1 day ago

Jina Releases Multimodal jina-embeddings-v5-omni-small

Jina has launched jina-embeddings-v5-omni-small, a multimodal model that searches video and audio across existing text indices without requiring...

AlphaSignal|4 days ago

TII Releases Falcon-ASR Arabic Speech Model

Abu Dhabi's Technology Innovation Institute introduced Falcon-ASR, a 1.6B parameter model that outperforms larger competitors on Arabic and regional...

MarkTechPost|1 day ago

OpenAI Launches Composer Predictions in Codex

OpenAI has introduced composer predictions in Codex, allowing the AI coding tool to anticipate and suggest a developer's next prompt during complex...

AlphaSignal|4 days ago

vLLM Semantic Router Releases Decision 3.0

The vLLM Semantic Router team has released Decision 3.0, five open multimodal decision models designed to route and classify requests via direct...

MarkTechPost|2 days ago

Nace AI Open-Sources Drex 1.5 Decision Model

Nace AI has open-sourced Drex 1.5, an 8.95-billion-parameter decision model that offers open-weights parity with proprietary alternatives for scoring...

MarkTechPost|3 days ago

Reflection AI and Mistral Release New Open-Weight Models

The launch of Reflection AI’s Beam and Mistral’s Large 4 preview gives Western enterprises powerful open-weight alternatives to proprietary systems...

AI Business|4 days ago

OrcaRouter Releases OrcaCyber Zero 1.5 Security Model

OrcaRouter has released OrcaCyber Zero 1.5, a security-focused model featuring a 1M-token context window designed for autonomous vulnerability...

MarkTechPost|2 days ago

PatientDesk AI Releases 30MB Antalia-2 Mini Speech Model

PatientDesk AI has launched Antalia-2 Mini, a 30.5 MB Turkish text-to-speech model that delivers benchmark-topping accuracy without needing expensive...

AlphaSignal|2 days ago

AutoTrust AI Launches 45ms GEV-26B Decision Model

AutoTrust AI Lab has launched GEV-26B-Decide, an open-weights decision API that delivers calibrated probability scores in just 45 milliseconds.

AlphaSignal|2 days ago

Google DeepMind Unifies Seven Tracking Models in TapNet

Google DeepMind has consolidated its complete Tracking Any Point ecosystem into a single open-source repository, offering unified code and weights...

AlphaSignal|3 days ago

ACE-Step 1.5 Brings Local Music Generation to C++

ACE-Step 1.5 is now available in quantized GGUF format alongside a lightweight C++17 runtime, enabling developers to generate 48kHz stereo music...

AlphaSignal|3 days ago

UkisAI Shrinks Swift 1.5 27B Model to 8.42GB

UkisAI has released mixed-precision GGUF quantizations of its Swift 1.5 27B reasoning model, compressing it to 8.42GB to enable local execution on...

AlphaSignal|1 day ago

SC117 Strips Refusals From Qwen3.8-Flash-Next

Developer SC117 has released an abliterated GGUF version of Qwen3.8-Flash-Next that strips safety refusals without degrading optimized low-bit...

AlphaSignal|2 days ago

Alibaba's 4-Bit Qwen2.5-Coder Hits 775K Downloads

A community 4-bit version of Alibaba's Qwen2.5-Coder-7B-Instruct has surpassed 775,000 downloads on Hugging Face, enabling developers to run capable...

AlphaSignal|4 days ago

Bartowski Shrinks Qwen2.5-32B Model Down to 9GB

Open-source developer Bartowski has published a suite of GGUF quantizations for Qwen2.5-32B, enabling local execution on consumer GPUs with memory...

AlphaSignal|1 day ago

DigUp Uses EmbeddingGemma 2 for Local Mac Search

A new free utility called DigUp brings local, cross-modal file search to Apple Silicon Macs using Google's EmbeddingGemma 2 model to query text,...

AlphaSignal|3 days ago

Hugging Face Adds RepVGG-A0 Model to timm

Hugging Face has integrated the pre-trained RepVGG-A0 model into its timm library, allowing developers to deploy an efficient 9.1-million-parameter...

AlphaSignal|3 days ago

Google Launches Playground AI to Generate Web Video Games

Google Labs has launched Playground, an experimental chatbot service that converts plain-text prompts into playable browser video games without...

WIRED AI|2 days ago

Umi-OCR Hits 47k Stars for Local Desktop Text Extraction

Open-source desktop tool Umi-OCR has reached roughly 47,000 GitHub stars, offering developers a completely offline, scriptable alternative to...

AlphaSignal|3 days ago

Simon Willison Updates ttok Tool for GPT-6 Tokenization

Developer Simon Willison has launched ttok 1.0, updating his command-line token utility to default to the latest GPT-5 and GPT-6 tokenizers for more...

Simon Willison|4 days ago