AllLLMCoding Co-pilotPersonal AssistantsAI AgentsHealthcare AIEthics & RegulationResearch & PapersStartups & Funding
Company:AllOpenAIAnthropicGoogleMicrosoftAppleDeepSeekMistralxAIMetaOthers

LLM

Large Language Models

Microsoft AI Introduces MAI-Transcribe-1.5: 2.4% WER on Artificial Analysis, Best-in-Class FLEURS Accuracy, and Up to 5x Faster Long-Audio Transcription
LLMMicrosoft

Microsoft AI Introduces MAI-Transcribe-1.5: 2.4% WER on Artificial Analysis, Best-in-Class FLEURS Accuracy, and Up to 5x Faster Long-Audio Transcription

Microsoft AI has released MAI-Transcribe-1.5, an improved speech-to-text model supporting 43 languages with 2.4%...

MarkTechPost
Building Reflective Prompt Optimization with GEPA: Multi-Component Prompts, Structured Feedback, and Held-Out Validation
LLM

Building Reflective Prompt Optimization with GEPA: Multi-Component Prompts, Structured Feedback, and Held-Out Validation

This tutorial demonstrates using GEPA, a reflective prompt-evolution framework, to optimize how small language models...

MarkTechPost
Deepseek topped Ramp's trending software vendors in June 2026 as US companies chase cheaper AI
LLMDeepSeek

Deepseek topped Ramp's trending software vendors in June 2026 as US companies chase cheaper AI

DeepSeek emerged as the top trending software vendor on Ramp in June 2026, driven by US companies seeking more...

The Decoder
Elon Musk's xAI reportedly trained its coding models on Claude outputs for months before getting cut off
LLMxAI

Elon Musk's xAI reportedly trained its coding models on Claude outputs for months before getting cut off

xAI trained its coding models using outputs from Anthropic's Claude for months, continuing even after access was cut...

The Decoder
New open-source voice model listens nonstop and decides every 0.4 seconds whether to speak or stay silent
LLM

New open-source voice model listens nonstop and decides every 0.4 seconds whether to speak or stay silent

A new open-source voice model called Audio Interaction enables continuous listening and can decide every 0.4 seconds...

The Decoder
NVIDIA Releases Nemotron 3.5 ASR: A 600M-Parameter Cache-Aware Streaming Model Transcribing 40 Language-Locales in Real Time
LLM

NVIDIA Releases Nemotron 3.5 ASR: A 600M-Parameter Cache-Aware Streaming Model Transcribing 40 Language-Locales in Real Time

NVIDIA released Nemotron 3.5 ASR, a 600M-parameter streaming model capable of transcribing 40 language-locales in real...

MarkTechPost
Google DeepMind Releases Gemma 4 QAT Checkpoints: Q4_0 and a New Mobile Format Cut On-Device Memory
LLMGoogle

Google DeepMind Releases Gemma 4 QAT Checkpoints: Q4_0 and a New Mobile Format Cut On-Device Memory

Google DeepMind has released Gemma 4 QAT checkpoints including Q4_0 and a new mobile format designed to reduce...

MarkTechPost
Google will pay SpaceX $920M per month for compute
LLMGoogle

Google will pay SpaceX $920M per month for compute

Google has agreed to pay SpaceX $920 million per month for compute resources to support unexpected demand for its AI...

TechCrunch AI
The token bill comes due: Inside the industry scramble to manage AI’s runaway costs
LLM

The token bill comes due: Inside the industry scramble to manage AI’s runaway costs

The AI industry is shifting focus from rapid expansion and cost maximization toward implementing cost controls and...

TechCrunch AI
NVIDIA AI Releases Nemotron 3 Ultra: An Open 550B Mixture-of-Experts Hybrid Mamba-Transformer for Long-Running Agents
LLM

NVIDIA AI Releases Nemotron 3 Ultra: An Open 550B Mixture-of-Experts Hybrid Mamba-Transformer for Long-Running Agents

NVIDIA has released Nemotron 3 Ultra, a 550B parameter Mixture-of-Experts hybrid Mamba-Transformer model designed for...

MarkTechPost
Google’s Gemma 4 12B Shows AI Race Moving to Edge Devices
LLMGoogle

Google’s Gemma 4 12B Shows AI Race Moving to Edge Devices

Google has released Gemma 4 12B under the Apache 2.0 license, demonstrating the industry trend of enabling enterprises...

AI Business
LLMOpenAI

Dreaming: Better memory for a more helpful ChatGPT

ChatGPT has introduced a new memory system that allows it to retain user preferences and maintain context across...

OpenAI Blog
Miso Labs Releases MisoTTS: An 8B Emotive Text-to-Speech Model with Open Weights
LLM

Miso Labs Releases MisoTTS: An 8B Emotive Text-to-Speech Model with Open Weights

Miso Labs has released MisoTTS, an open-weights 8 billion parameter text-to-speech model that uses residual vector...

MarkTechPost
xAI updates Grok Imagine to 1.5 with image-to-video generation at 720p resolution
LLMxAI

xAI updates Grok Imagine to 1.5 with image-to-video generation at 720p resolution

xAI has released Grok Imagine Video 1.5, an image-to-video generation model that converts still images into cinematic...

The Decoder
Google Deepmind's Gemma 4 12B squeezes multimodal AI onto a laptop with just 16 GB of RAM
LLMGoogle

Google Deepmind's Gemma 4 12B squeezes multimodal AI onto a laptop with just 16 GB of RAM

Google DeepMind released Gemma 4 12B, an open-source multimodal AI model that processes text, images, and audio...

The Decoder
Google DeepMind Releases Gemma 4 12B: An Encoder-Free Multimodal Model with Native audio that runs on a 16 GB laptop
LLMGoogle

Google DeepMind Releases Gemma 4 12B: An Encoder-Free Multimodal Model with Native audio that runs on a 16 GB laptop

Google DeepMind has released Gemma 4 12B, an encoder-free multimodal model that natively processes vision and audio...

MarkTechPost
Ideogram 4.0 drops as an open-weight model with native 2K resolution and improved text rendering
LLM

Ideogram 4.0 drops as an open-weight model with native 2K resolution and improved text rendering

Ideogram has released version 4.0 of its text-to-image model as an open-weight model featuring native 2K resolution,...

The Decoder
Perplexity announces hybrid AI system that decides what runs locally or in the cloud
LLM

Perplexity announces hybrid AI system that decides what runs locally or in the cloud

Perplexity has announced a hybrid AI orchestrator system that intelligently distributes tasks between local and...

The Decoder
Build 2026: Microsoft tops Google in image generation while playing catch-up on reasoning
LLMMicrosoft

Build 2026: Microsoft tops Google in image generation while playing catch-up on reasoning

Microsoft announced seven new in-house AI models at Build 2026, including its first reasoning model, a new tuning...

The Decoder