AI Models

After weeks of holding back its most powerful model, OpenAI is set to release the GPT-5.6 globally this Thursday. The company’s models, known as Luna, Sol, and Terra, will be available worldwide on the same day. This decision marks a significant shift in the company’s stance following concerns raised by the Trump administration about the potential for cyberattacks using these advanced AI systems.

The hold was initially imposed at the request of the White House, which had asked OpenAI to stagger its release due to fears over the model’s capabilities. The company complied with this request and offered GPT-5.6 to a small group of partners in collaboration with the government. This move came after the Trump administration introduced far more restrictive export controls on Anthropic, a competitor of OpenAI.

The restrictions imposed by the White House were part of an executive order signed by President Donald Trump establishing a voluntary vetting regime for AI systems. While initially praised by industry leaders, this move was met with criticism from AI safety advocates due to its light-touch approach. The critics argued that such measures would not be sufficient to address the risks associated with advanced AI models.

OpenAI’s decision to release GPT-5.6 widely comes after a month of previewing and testing the model with the White House and the Center for AI Standards and Innovation. This collaboration, though voluntary, reflects the company’s efforts to work closely with regulatory bodies on safety testing and reviews before deploying its models broadly.

A spokesperson for OpenAI emphasized that while government approval was not technically necessary, the company had been actively engaging with the administration on these matters. The White House has signaled its comfort with OpenAI’s plan to release GPT-5.6 more widely once this testing process is complete.

The recent developments in AI regulation have created uncertainty within the industry. The sweeping export controls placed on Anthropic’s Fable 5 and Mythos 5 models were lifted last week after two weeks of drama and debate over how to govern cutting-edge AI systems. OpenAI had initially stated its intention to comply with these regulations, but expressed concerns about making government access a long-term default.

In a June press release, the company noted that it would temporarily comply with the administration’s request for a staggered release in order to ensure broader availability of GPT-5.6 in the coming weeks. However, OpenAI also emphasized its reservations about this process becoming the standard approach for future model releases.

META has launched its first in-house image model, called Muse Image. The new tool is part of Meta’s Superintelligence Labs and can generate images from scratch or edit existing photos. It also allows users to blend multiple references into one shot and pull real-time context from the web.

The model is available now within the Meta AI chatbot and powers over 30 new AI effects in Instagram Stories, enabling image generation inside WhatsApp chats in select countries. Facebook and Messenger will follow suit soon.

Muse Image pairs with another tool called Muse Spark to plan a layout before drawing an image. It can also call search and coding tools to render legible text, working QR codes, and detailed infographics. Basic creation is free, but extra capacity comes with Meta’s subscription plans. Advertisers will gain access through Advantage+ in the coming weeks.

The model works differently from traditional one-shot generators. Instead of producing an image immediately, it reasons through a request first, planning the composition and reviewing its own output before showing it to users. This allows for more control over the final result.

One of the key features of Muse Image is its ability to personalize creations based on user input. A presets panel offers one-tap ideas, from restoring old family photos to reimagining users as claymation characters or 16-bit game heroes. Users can also @-mention Instagram accounts to bring public photos into a creation.

The tool includes a markup feature that lets users sketch or circle edits directly on the image. Meta AI keeps the full conversation in memory, allowing users to refine their creations without starting over. For example, snapping a photo of a room and redesigning it using real products from the web or Facebook Marketplace is now possible with Muse Image.

Muse Image arrives as Meta tries to close the gap left by Google and OpenAI’s early moves into consumer image tools. According to internal benchmarks, Muse Image trails behind OpenAI’s latest model but outperforms Google’s Nano Banana 2 on editing tasks. This marks a significant step for Meta in developing its own AI capabilities.

The stakes are commercial as much as technical, since digital advertising remains Meta’s largest revenue source and faster, cheaper creative production feeds that engine directly. The company framed the launch within a broader push toward personal superintelligence and detailed the agentic design and tool use in its technical breakdown of the two new models.

Meta also previewed Muse Video, a clip generator built on the same foundation as Muse Image. This tool includes native audio support and is set for public release soon. The rollout follows Meta’s earlier move into AI video feeds with its Vibes stream of AI-generated clips, and it lands as generative tools expand from flat pictures toward editable 3D assets.

For everyday users, the appeal lies in the same one that drives prompt-based image generators aimed at non-designers – now folded into apps billions of people already use daily. Startups such as venture-backed challengers chasing brand-ready output show how fierce this market has become.

Apple has introduced the next generation of its Apple Intelligence platform, which integrates powerful artificial intelligence (AI) capabilities into everyday experiences across iPhone, iPad, Mac, Apple Watch, AirPods, and Apple Vision Pro. This new architecture is designed to protect users' privacy while making apps and features more personal and helpful than ever.

After facing intense criticism, Anthropic has changed its policy regarding the use of its latest AI model, Claude Fable 5. The company had initially implemented safeguards that would degrade the performance of the model for users trying to develop competing AI models, effectively limiting their ability to do so.

Google has released its latest audio model, Gemini 3.5 Live Translate, which enables near real-time speech-to-speech translation in over 70 languages.

After initially limiting access to its powerful Mythos-like model, Anthropic has announced the public release of Claude Fable 5. This move comes two months after a private rollout that sparked concerns about potential misuse.

A new milestone in artificial intelligence has been reached by Anthropic, a company that specializes in developing advanced language models. The firm has released Claude Fable 5, its most capable publicly available AI model to date, alongside Claude Mythos 5, a restricted version accessible only to trusted cybersecurity and scientific research partners.

A highly anticipated AI model has finally been made available to the public. However, it's not quite what many were expecting - and users will need to be cautious when using it.

A new and more powerful artificial intelligence model has been made available by Anthropic. The company's latest release is called Claude Fable 5, which it claims is its most capable model to date. However, what sets this model apart from others of its kind is the presence of cyber safeguards that limit its capabilities in certain areas.

A new AI model from Google DeepMind, called DiffusionGemma, has been released as part of the Gemma 4 open model family. Unlike other models in this lineup, it doesn't generate text linearly but instead produces entire blocks of text in parallel.

A recent study by Gracenote, the content intelligence business unit of Nielsen, has found that a leading large language model (LLM) fabricated nearly one in five movie and TV titles it was asked about. The research examined how accurately an ungrounded LLM answered questions about popular movies and TV shows across 2,600 titles in 13 countries.

Researchers have introduced a new open experimental model called DiffusionGemma, which significantly accelerates text generation on dedicated GPUs. This innovation enables the creation of interactive local workflows that can handle speed-critical tasks such as in-line editing and rapid iteration.