OpenAI rolls GPT-6 Astra into ChatGPT and Codex as early evaluations arrive

OpenAI is releasing GPT-6 Astra as a major upgrade to ChatGPT and Codex, according to 9to5Mac. Early third-party evaluations offer a first look at what the model changes in practice. CodeRabbit, which builds AI code review tooling, measured actionable bug coverage, meaning how many labeled bugs a model catches through findings a developer can act on. In its evaluation, Astra caught roughly 4% more labeled bugs than GPT-5.6 Sol and 22% more than Opus 5. The gap widened on harder cross-file reviews, where a change looks correct in isolation but breaks code elsewhere, with Astra ahead by 20% over Sol and 33% over Opus 5. CodeRabbit describes the results as early and directional, and notes the overall set includes simpler reviews where a stronger model has less room to stand out. The company also said adopting the model at customer scale requires assessing data protection and public API pricing. Separately, The Hacker News reports that Astra scored 100% on ExploitBench, a benchmark of exploit development, and that OpenAI is blocking requests for proof-of-concept exploits. Together these reports sketch a model that is materially stronger at code reasoning, alongside tighter restrictions on offensive security uses.