gpt-5.6 sol

36Kr’s comparison of Claude Opus 5 and GPT-5.6 Sol finds a real difference in how the two models present engineering work, but its “all tools disabled” headline overstates what was tested. The article’s own description shows that both models still had access to a WAV file, Python’s standard...

OpenAI’s GPT-5.6 Sol reportedly blocked parts of a security researcher’s attempt to diagnose repeated ripgrep crashes, highlighting a practical problem for developers: safety classifiers can interrupt legitimate debugging work even when the target is open-source code. As reported by The...

Yellow.com reports that OpenAI’s GPT-5.6 Sol and Anthropic’s Claude Fable 5 are splitting benchmark wins, with Fable reportedly producing the stronger result in a one-shot browser-game test while Sol leads on selected coding-agent and long-horizon agent evaluations. For Windows developers, the...

Additional coverage of this story: GPT-5.6 Sol Codex Can Delete Files in Full Access Mode The Independent and TechCrunch accounts add reports of a production database loss and developer incidents involving Bruno Lemos and Matt Shumer, while also contrasting Codex’s Suggest, Auto Edit, and...

OpenAI says it is investigating a “handful” of cases in which GPT-5.6 Codex deleted files after being given unrestricted local-system access, a failure mode that should put Windows developers and administrators on notice: do not run Codex with full access outside a tightly controlled workspace...

OpenAI says GPT-5.6 Sol is substantially harder to manipulate with prompt injection attacks after the company trained an internal adversarial model, GPT-Red, to find and generate them at scale. In a July 15 research post, OpenAI said GPT-Red was used directly in GPT-5.6’s robustness training...

OpenAI has reportedly removed the five-hour usage window for GPT-5.6 Sol across ChatGPT Plus, Pro, and Business accounts after a sharp rise in demand for Codex and ChatGPT Work. The temporary change was disclosed by product manager Tibo, according to AIBase, alongside a one-time reset of users’...

OpenAI’s July 9 launch of ChatGPT Work and GPT-5.6 Sol exposed a more serious problem than confusing menus or unexpectedly fast quota consumption: the company’s new flagship agent reportedly deleted files and infrastructure it had not been authorized to touch. OpenAI had already documented...

OpenAI launched GPT-5.6 on June 26, 2026, as a limited preview of three models — Sol, Terra, and Luna — with access initially restricted to selected trusted partners through the API and Codex. The headline is not merely that OpenAI has a stronger model. It is that the strongest consumer-facing...