Sol, Terra and Luna appear differently across ChatGPT, Work, Codex and the API. A guide to choosing by task and measuring cost per result, not isolated token price.
An agent is not total autonomy but an allocation of control. Goal, tools, verification, authority and evidence make the claim auditable.
Apple accuses OpenAI, its hardware chief Tang Tan, engineer Chang Liu and io Products of misappropriating iPhone trade secrets to accelerate its first consumer device. The federal lawsuit, filed in San Jose, brings six claims and seeks injunctions and damages; OpenAI replies that it has "no interest in other companies' trade secrets." The case will test where the line sits between hiring talent and taking confidential know-how in the middle of the AI hardware race.
OpenAI expands access to Sol, Terra and Luna and introduces ChatGPT Work, a surface that gathers context and acts through tools. Evaluating the agent requires separating the model, permissions, actions and verification.
The memo is not public, and “code red” does not mean defeat. A guide to separating facts, attribution, incompatible metrics and still-pending consequences.
Kimi K2 Thinking combines downloadable weights, mixture-of-experts routing, quantization, and long tool use. Evaluating it means separating memory from active compute, licensing from reproducibility, and call count from useful outcomes.
OpenAI’s browser can read pages, remember context, and act on websites, but those capabilities do not carry the same risk. A matrix of data, permissions, and confirmations makes delegation possible without handing over the entire session.
OpenAI brought conversational apps, agent tools, and new API models together at DevDay. The useful skill is separating interface, data, action, authorization, and evaluation before entrusting a task.
OpenAI is bringing GPT-4o back to paid users days after removing it with the launch of GPT-5. The backlash shows that choosing a model is no longer just about speed or capability.
OpenAI is making GPT-5 ChatGPT’s default model, including for free accounts. The system decides when to answer instantly and when to spend more compute on reasoning, with announced improvements in coding, healthcare and reliability.
OpenAI releases gpt-oss-120b and gpt-oss-20b under the Apache 2.0 license. They are its first open-weight language models since GPT-2 and can run reasoning capabilities on users’ own infrastructure.
OpenAI brings Operator and Deep Research together in ChatGPT Agent, a system that browses the web, analyzes data and uses a virtual computer to complete tasks. The launch brings autonomous agents closer to the mainstream, with limitations and risks the company acknowledges.
This website uses cookies to improve the browsing experience. Cookie policy.
No results. Try different words.
