OpenAI launches most capable GPT model after delayed rollout
The rollout includes three model tiers, new ChatGPT Work tools and pricing from $1 to $30 per million tokens, OpenAI said.
- On Tuesday, OpenAI publicly launched GPT-5.6, its most advanced AI model, following a delay last month prompted by U.S. government national security concerns about potential misuse.
- President Donald Trump signed an executive order establishing a voluntary framework requiring AI developers to provide "covered frontier models" to the U.S. government for up to 30 days before release.
- GPT-5.6 splits into three tiers: Sol for demanding tasks, Terra as a "balanced" option, and Luna for cost-efficient everyday use, priced from $1 to $5 input per one million tokens.
- OpenAI introduced ChatGPT Work, an agent managing files and apps, available immediately to desktop users including Free subscribers, with expansion to Plus and Business users "over the next few days."
- Billionaire Elon Musk announced his company's leading model, Grok 4.5, is now available to the public, intensifying competition as Anthropic warned it is "probably impossible" to make models fully robust against jailbreaks.
13 Articles
13 Articles
OpenAI launches most capable GPT model after delayed rollout
OpenAI has publicly launched GPT-5.6, its most advanced AI model, following a delay last month prompted by US government requests over heightened national security concerns about the potential misuse of powerful AI technologies. The United States is in a race to develop cutting-edge AI models the likes of which, experts have said, could dramatically accelerate […]
OpenAI presented yesterday GPT-5.6, a family of three models, together with ChatGPT Work, a productivity agent that combines ChatGPT with Codex to perform office tasks without programming knowledge. The launch is the company’s most openly competitive in months: benchmarks are compared directly with Anthropic, and the nomenclature of ChatGPT Work mimics Claude Cowork, the agent that Anthropic launched in January 2026. Lucas Ropek publishes on Tec…
The update introduces improvements in understanding, reasoning and follow-up instructions to provide more useful and consistent responses.
OpenAI's GPT-5.6 Sol autonomously post-trained the smaller Luna model with a "fairly underspecified prompt"
According to OpenAI, GPT-5.6 Sol independently fine-tuned the smaller Luna model, triggered by a single "fairly under-specified prompt." In OpenAI's internal RSI benchmark for recursive self-improvement, Sol scores 16.2 points higher than GPT-5.5. OpenAI believes the "automated researcher" is within reach. The article OpenAI's GPT-5.6 Sol autonomously post-trained the smaller Luna model with a "fairly underspecified prompt" appeared first on The…
According to the company, OpenAI's GPT-5.6 Sol has independently post-trained the smaller Luna model – initiated by a single, "quitely underspecified prompt". In the internal RSI benchmark for recursive self-improvement, Sol is 16.2 points ahead of GPT-5.5. OpenAI sees the "automated researcher" within reach. The article AI that improves AI: OpenAI's GPT-5.6 Sol supposedly trained smaller Luna model appeared first on The Decoder.
Coverage Details
Bias Distribution
- 50% of the sources lean Left, 50% of the sources lean Right
Factuality
To view factuality data please Upgrade to Premium
