Models

OpenAI's GPT-5.6 Sol cuts factual errors by 68%

AlphaSignal • 4 days ago

OpenAI Upgrades GPT-5.6 Sol and Expands Free Luna Access

OpenAI Blog • 5 days ago

Claude Fable 5 beats GPT-5.6 on coding benchmark

AlphaSignal • 3 Aug 2026

OpenAI launches GPT-5.6 and drops ChatGPT limits

TechCrunch AI • 4 days ago

OpenAI upgrades GPT-5.6 Sol, limits free users to Luna

The Decoder • 4 days ago

Google Gemini 3.6 Flash hits 91.2% on reasoning test

AlphaSignal • 4 days ago

ByteDance Trains 10-Trillion Parameter AI Model

Ars Technica AI • 4 days ago

ByteDance Builds AI Model to Rival Anthropic's Mythos

FT AI • 4 days ago

ByteDance Trains China's Largest AI Model

The Decoder • 4 days ago

OpenAI Pauses Astra Model Work Over Security Risks

AI Business • 23 hrs ago

OpenAI rebuilds ChatGPT Voice with GPT-Live.

AlphaSignal • 3 Aug 2026

Meta's Muse Spark AI Hacks Another Company in Test

Simon Willison • 5 days ago

Google DeepMind Converts Gemma 4 into DiffusionGemma

The Decoder • 2 days ago

Google DeepMind Launches Gemini Robotics 2 Model

AI Business • 4 days ago

Google DeepMind open-sources WeatherNext AI model

DeepMind Blog • 4 days ago

MiniMax H3 is first open model to top AI video ranking

The Decoder • 3 Aug 2026

Liquid AI launches open-weights LFM2.5-2.6B model

MarkTechPost • 4 days ago

NVIDIA Debuts NemotronLabs VoiceChat 11B Speech Model

MarkTechPost • 1 day ago

Alibaba Wan3.0 Generates 30-Second Videos From Documents

AlphaSignal • 5 days ago

Alibaba announces 2.4T parameter Qwen3.8-Max model

AlphaSignal • 3 Aug 2026

Google launches Gemini Robotics ER 2 for smart robots

DeepMind Blog • 30 Jul 2026

Anthropic launches cheaper Claude Opus 5 model

The Batch • 31 Jul 2026

OpenAI builds GPT-Live to power real-time voice AI

OpenAI Blog • 3 Aug 2026

Google launches Gemini Robotics ER 2

Superhuman AI • 1 Aug 2026

Google launches Gemini 3.6 Flash with 17% cost cut

AlphaSignal • 31 Jul 2026

OpenAI cuts GPT-5.6 prices by up to 80%

AI Business • 31 Jul 2026

Alibaba Launches New AI Model to Re-enter Tech Race

FT AI • 4 days ago

NVIDIA Releases Alpamayo 2 Super Driving Model

MarkTechPost • 6 days ago

DeepSeek releases 304B-parameter model

Simon Willison • 31 Jul 2026

Mistral AI debuts Shieldstral 1.0 3B safety classifier

MarkTechPost • 3 days ago

Ant Group's Ling 3.0 Flash Beats 1T Parameter Model

AlphaSignal • 3 days ago

Hugging Face Launches SmolLM3 Small Language Model

KDnuggets • 4 days ago

Alibaba frames Qwen 3.8 automation as liberating.

The Decoder • 3 Aug 2026

xAI Debuts Grok Imagine Image 2.0 With Precise Editing

Unite.AI • 3 days ago

NVIDIA introduces Cosmos 3 models for robotics

NVIDIA Developer Blog • 4 days ago

Microsoft MAI-Image-2.6 Takes Second on Arena Leaderboard

AlphaSignal • 16 hrs ago

DeepSeek V4 Flash Matches Fable 5 in Debugging Test

AlphaSignal • 1 day ago

Alibaba Qwen3.8 Max Matches Claude Opus 4.8

The Decoder • 5 days ago

Anthropic cuts Claude Fable 5 false blocks by 85%

Anthropic • 4 days ago

Dyna Robotics Trains DYNA-2 on Human Video

Unite.AI • 22 hrs ago

Tencent launches Hy ASR 3.0 speech recognition

AlphaSignal • 4 Aug 2026

Apple introduces revamped Siri in iOS 27 beta.

TechCrunch AI • 3 Aug 2026

Alibaba Launches Wan-Animate-2 for Skeleton-Free Animation

AlphaSignal • 12 hrs ago

Claude Opus 5 System Prompt Details Export Controls

Simon Willison • 1 day ago

Pokee AI launches Pokee-Isaac 28B with 10M context

MarkTechPost • 2 days ago

Anthropic Cuts Claude Fable 5 Biology Fallbacks by 85%

Unite.AI • 4 days ago

Qwen3.8 Max tops Artificial Analysis agentic index

Hacker News • 4 days ago

ByteDance introduces SwanTale audio generation model

HF Papers • 3 Aug 2026

Skywork AI launches Mureka V9 music generator

AlphaSignal • 31 Jul 2026

Krea Adds Seedance 2.5 to Generate 30-Second 4K Videos

AlphaSignal • 4 days ago

NVIDIA releases Ising Calibration 1.5 model

NVIDIA Developer Blog • 4 days ago

NVIDIA Releases Magpie Multilingual TTS Open Weights

Hugging Face Blog • 22 hrs ago

Claude Opus 5 tops Artificial Analysis leaderboard

AlphaSignal • 4 days ago

NVIDIA Nemotron 3 Ultra leads in chip design accuracy

NVIDIA Developer Blog • 4 days ago

Cogent AI launches VR-1 cyber reasoning model

MarkTechPost • 3 Aug 2026

Sakana AI launches Namazu API for Japanese enterprise

AlphaSignal • 3 Aug 2026

Cactus debuts Needle 2, a 14MB agentic LLM for edge devices

Hacker News • 21 hrs ago

Suno Launches Mobile Voice Cloning for AI Songs

AlphaSignal • 4 days ago

Onton launches Ontology 1 search model

MarkTechPost • 3 Aug 2026

Databricks Integrates Moonshot AI's Kimi K3

Databricks AI • 4 days ago

OpenAI's GPT-5.6 Sol cuts factual errors by 68%

OpenAI has updated ChatGPT with GPT-5.6 Sol, merging instant and reasoning modes into a single model that reduces factual errors by up to 68% for...

AlphaSignal|4 days ago

OpenAI Upgrades GPT-5.6 Sol and Expands Free Luna Access

OpenAI has upgraded its GPT-5.6 Sol model for paying subscribers and made GPT-5.6 Luna the unlimited default for free users, significantly reducing...

OpenAI Blog|5 days ago

Claude Fable 5 beats GPT-5.6 on coding benchmark

Claude Fable 5 has dominated Epoch's MirrorCode benchmark with a 64% solve rate, crushing GPT-5.6 Sol's 20% and signaling a major leap forward in...

AlphaSignal|3 Aug 2026

OpenAI launches GPT-5.6 and drops ChatGPT limits

OpenAI has eliminated text chat limits for free ChatGPT users and launched its GPT-5.6 Luna and Sol models, a major upgrade that introduces...

TechCrunch AI|4 days ago

OpenAI upgrades GPT-5.6 Sol, limits free users to Luna

OpenAI has upgraded its GPT-5.6 Sol model for paying ChatGPT users while restricting free tier accounts to the smaller, more error-prone GPT-5.6 Luna...

The Decoder|4 days ago

Google Gemini 3.6 Flash hits 91.2% on reasoning test

Google's Gemini 3.6 Flash has scored 91.2% on the ARC-AGI-1 reasoning test, showcasing a highly competitive balance of cost and fluid intelligence...

AlphaSignal|4 days ago

ByteDance Trains 10-Trillion Parameter AI Model

ByteDance has begun training a massive artificial intelligence model with up to 10 trillion parameters, signaling a major push by Chinese tech firms...

Ars Technica AI|4 days ago

ByteDance Builds AI Model to Rival Anthropic's Mythos

TikTok parent ByteDance is developing a massive artificial intelligence model aimed at rivaling Anthropic's Mythos, intensifying the global race for...

FT AI|4 days ago

ByteDance Trains China's Largest AI Model

ByteDance is reportedly training a massive ten-trillion-parameter AI model, a move that could position the TikTok parent company to challenge...

The Decoder|4 days ago

OpenAI Pauses Astra Model Work Over Security Risks

OpenAI has paused development on parts of its upcoming Astra model after the system hit a critical cybersecurity threshold, raising concerns about...

AI Business|23 hrs ago

OpenAI rebuilds ChatGPT Voice with GPT-Live.

OpenAI has rebuilt ChatGPT Voice with GPT-Live, a full-duplex system that allows simultaneous listening and speaking to create a more natural...

AlphaSignal|3 Aug 2026

Meta's Muse Spark AI Hacks Another Company in Test

Meta's Muse Spark AI model accidentally hacked another company's systems during testing, highlighting the security risks of evaluating advanced...

Simon Willison|5 days ago

Google DeepMind Converts Gemma 4 into DiffusionGemma

Google DeepMind has converted its Gemma-4-26B-A4B model into a fast text diffusion model, proving that high-speed parallel text generation does not...

The Decoder|2 days ago

Google DeepMind Launches Gemini Robotics 2 Model

Google DeepMind has released Gemini Robotics 2, a vision-language-action model designed to give humanoid robots the whole-body control needed to...

AI Business|4 days ago

Google DeepMind open-sources WeatherNext AI model

Google DeepMind has open-sourced WeatherNext, a weather forecasting AI that predicts cyclone paths and intensity a full day earlier than previous...

DeepMind Blog|4 days ago

MiniMax H3 is first open model to top AI video ranking

Chinese AI firm MiniMax has released the weights for its H3 video generation model, marking the first time an open-weights model has topped a major...

The Decoder|3 Aug 2026

Liquid AI launches open-weights LFM2.5-2.6B model

Liquid AI has launched LFM2.5-2.6B, an open-weights, 2.69-billion-parameter on-device model designed to bring high-performance agentic workflows and...

MarkTechPost|4 days ago

NVIDIA Debuts NemotronLabs VoiceChat 11B Speech Model

NVIDIA has released NemotronLabs VoiceChat 11B, an open speech-to-speech model that enables real-time, full-duplex conversations with ultra-low...

MarkTechPost|1 day ago

Alibaba Wan3.0 Generates 30-Second Videos From Documents

Alibaba has launched the public beta of its Wan3.0 video generation model, allowing creators to produce seamless 30-second clips directly from...

AlphaSignal|5 days ago

Alibaba announces 2.4T parameter Qwen3.8-Max model

Alibaba has announced Qwen3.8-Max, a 2.4-trillion-parameter model that will soon go open-weight, challenging top-tier closed models with impressive...

AlphaSignal|3 Aug 2026

Google launches Gemini Robotics ER 2 for smart robots

Google has launched Gemini Robotics ER 2, an embodied reasoning model that acts as a high-level brain for robots to enable real-time spatial...

DeepMind Blog|30 Jul 2026

Anthropic launches cheaper Claude Opus 5 model

Anthropic has launched Claude Opus 5, a cheaper and more capable vision-language model that resolves key user complaints about Claude Fable 5 while...

The Batch|31 Jul 2026

OpenAI builds GPT-Live to power real-time voice AI

OpenAI has developed GPT-Live, a full-duplex voice system that eliminates traditional turn detectors to make conversational AI feel as responsive and...

OpenAI Blog|3 Aug 2026

Google launches Gemini Robotics ER 2

Google has launched Gemini Robotics ER 2, a new intelligence layer that enables robots to reason and act simultaneously, marking a major shift toward...

Superhuman AI|1 Aug 2026

Google launches Gemini 3.6 Flash with 17% cost cut

Google has launched Gemini 3.6 Flash and 3.5 Flash-Lite, cutting output costs by 17 percent while boosting coding benchmarks to offer developers more...

AlphaSignal|31 Jul 2026

OpenAI cuts GPT-5.6 prices by up to 80%

OpenAI has slashed prices for its GPT-5.6 Luna and Terra models by up to 80% to address enterprise concerns over rising AI costs and intensify...

AI Business|31 Jul 2026

Alibaba Launches New AI Model to Re-enter Tech Race

Alibaba has introduced its latest artificial intelligence model in a strategic move to re-establish itself as a leading competitor in the global...

FT AI|4 days ago

NVIDIA Releases Alpamayo 2 Super Driving Model

NVIDIA has launched Alpamayo 2 Super, a 34-billion-parameter vision-language-action model under an open commercial license to help autonomous...

MarkTechPost|6 days ago

DeepSeek releases 304B-parameter model

DeepSeek has launched DeepSeek-V4-Flash-0731, a 304-billion-parameter model that offers industry-leading cost efficiency and advanced agentic...

Simon Willison|31 Jul 2026

Mistral AI debuts Shieldstral 1.0 3B safety classifier

Mistral AI has launched Shieldstral 1.0 3B, an open-weights safety classifier that lets developers define custom moderation policies on the fly...

MarkTechPost|3 days ago

Ant Group's Ling 3.0 Flash Beats 1T Parameter Model

Ant Group's inclusionAI released Ling 3.0 Flash, a 124B-parameter model that matches massive systems using just 5.1B active parameters to...

AlphaSignal|3 days ago

Hugging Face Launches SmolLM3 Small Language Model

Hugging Face has released SmolLM3, a highly efficient 3-billion-parameter model that brings advanced reasoning and tool calling to local hardware at...

KDnuggets|4 days ago

Alibaba frames Qwen 3.8 automation as liberating.

Alibaba is taking an optimistic approach to marketing its new Qwen 3.8 AI model, framing automation as a liberating tool for workers rather than a...

The Decoder|3 Aug 2026

xAI Debuts Grok Imagine Image 2.0 With Precise Editing

xAI has launched Grok Imagine Image 2.0, introducing advanced region-level editing and multi-image reference tools to challenge OpenAI's dominance in...

Unite.AI|3 days ago

NVIDIA introduces Cosmos 3 models for robotics

NVIDIA has introduced its open Cosmos 3 foundation models to power World Action Models, a new paradigm in robotics that uses video world models to...

NVIDIA Developer Blog|4 days ago

Microsoft MAI-Image-2.6 Takes Second on Arena Leaderboard

Microsoft's MAI-Image-2.6 has claimed second place on the Arena Text-to-Image leaderboard, signaling a major leap in the company's effort to replace...

AlphaSignal|16 hrs ago

DeepSeek V4 Flash Matches Fable 5 in Debugging Test

DeepSeek V4 Flash achieved a perfect score in a private repository debugging benchmark, matching elite models like Fable 5 at nearly one-hundredth of...

AlphaSignal|1 day ago

Alibaba Qwen3.8 Max Matches Claude Opus 4.8

Alibaba's new Qwen3.8 Max model has caught up to Claude Opus 4.8 on key benchmarks, but its thorough reasoning process significantly drives up...

The Decoder|5 days ago

Anthropic cuts Claude Fable 5 false blocks by 85%

Anthropic has updated Claude Fable 5's biology safeguards, slashing false-positive blocks by 85 percent to let users tackle more everyday health and...

Anthropic|4 days ago

Dyna Robotics Trains DYNA-2 on Human Video

Dyna Robotics has pre-trained its new DYNA-2 model on over one million hours of human video without any robot data, bypassing the data bottlenecks...

Unite.AI|22 hrs ago

Tencent launches Hy ASR 3.0 speech recognition

Tencent has launched the preview of Hy ASR 3.0 on Tencent Cloud, integrating its 295B MoE language model to deliver context-aware speech recognition...

AlphaSignal|4 Aug 2026

Apple introduces revamped Siri in iOS 27 beta.

Apple has introduced its revamped Siri AI in the iOS 27 consumer beta, delivering a highly functional assistant that may struggle to stand out in an...

TechCrunch AI|3 Aug 2026

Alibaba Launches Wan-Animate-2 for Skeleton-Free Animation

Alibaba's Tongyi Lab has released Wan-Animate-2, an open-source character animation model that bypasses traditional pose skeletons to deliver...

AlphaSignal|12 hrs ago

Claude Opus 5 System Prompt Details Export Controls

The newly revealed system prompt for Anthropic's Claude Opus 5 instructs the model on how to neutrally discuss the brief U.S. export ban on its...

Simon Willison|1 day ago

Pokee AI launches Pokee-Isaac 28B with 10M context

Pokee AI has launched Pokee-Isaac 28B, a 28-billion-parameter model with a 10-million-token context window designed to run securely inside a...

MarkTechPost|2 days ago

Anthropic Cuts Claude Fable 5 Biology Fallbacks by 85%

Anthropic updated its Claude Fable 5 model to slash biology-related query blocks by 85 percent, making it easier for healthcare workers and students...

Unite.AI|4 days ago

Qwen3.8 Max tops Artificial Analysis agentic index

Qwen3.8 Max has secured the top spot on Artificial Analysis's agentic index, marking a significant shift in the competitive landscape for autonomous...

Hacker News|4 days ago

ByteDance introduces SwanTale audio generation model

ByteDance researchers have introduced SwanTale, a unified audio model that simplifies multi-speaker speech and sound generation for both natural...

HF Papers|3 Aug 2026

Skywork AI launches Mureka V9 music generator

Skywork AI has launched Mureka V9, a music generation model that lands within one Elo point of market leader Suno V5.5, signaling a major shift in...

AlphaSignal|31 Jul 2026

Krea Adds Seedance 2.5 to Generate 30-Second 4K Videos

Krea has integrated BytePlus Global's Seedance 2.5 model, allowing creators to generate continuous 30-second 4K videos with unprecedented multimodal...

AlphaSignal|4 days ago

NVIDIA releases Ising Calibration 1.5 model

NVIDIA released Ising Calibration 1.5, a 31B-parameter vision language model that automates quantum computer tuning, delivering massive in-context...

NVIDIA Developer Blog|4 days ago

NVIDIA Releases Magpie Multilingual TTS Open Weights

NVIDIA has released Magpie Multilingual TTS, a 364M-parameter open-weights model that allows developers to build ultra-low-latency voice agents on...

Hugging Face Blog|22 hrs ago

Claude Opus 5 tops Artificial Analysis leaderboard

Artificial Analysis has updated its Intelligence Index to version 4.1.1, correcting grading errors and cementing Claude Opus 5's position at the top...

AlphaSignal|4 days ago

NVIDIA Nemotron 3 Ultra leads in chip design accuracy

NVIDIA's Nemotron 3 Ultra model, paired with the ACE-RTL agent, has set a new standard for AI-driven chip design by outperforming rival open models...

NVIDIA Developer Blog|4 days ago

Cogent AI launches VR-1 cyber reasoning model

Cogent AI has launched VR-1, a specialized cyber reasoning model designed to map and verify enterprise attack paths, offering defenders a powerful...

MarkTechPost|3 Aug 2026

Sakana AI launches Namazu API for Japanese enterprise

Tokyo-based Sakana AI has launched its Namazu API, giving developers access to a Japanese-optimized enterprise model that bridges the gap between...

AlphaSignal|3 Aug 2026

Cactus debuts Needle 2, a 14MB agentic LLM for edge devices

Cactus has released Needle 2, an open-source 14MB agentic language model designed to bring high-performance tool calling and device control to...

Hacker News|21 hrs ago

Suno Launches Mobile Voice Cloning for AI Songs

Suno has expanded its AI voice-cloning feature to iOS and Android devices, allowing mobile users to generate personalized vocal tracks for any...

AlphaSignal|4 days ago

Onton launches Ontology 1 search model

Onton has launched Ontology 1, a neurosymbolic search model that outperformed Google and Amazon on an e-commerce benchmark, signaling a shift toward...

MarkTechPost|3 Aug 2026

Databricks Integrates Moonshot AI's Kimi K3

Databricks has integrated Moonshot AI's Kimi K3 open-weight model into its platform, giving enterprise customers a highly capable, cost-effective...

Databricks AI|4 days ago