Users are expressing disappointment with OpenAI's GPT-6 Astra model, claiming its performance has declined just a week after its launch.
Backlash Grows Against OpenAI's GPT-6 Astra Model After Launch
Just a week post-launch, OpenAI's GPT-6 Astra is facing significant criticism from users who claim that its performance has noticeably declined. Initially celebrated for its impressive capabilities, including the ability to recreate complex environments in game engines, Astra's perceived drop in quality has sparked a wave of discontent on social media.
Developer synthwavedd expressed disappointment on X, stating, "Astra feels significantly dumber for me today. It was only a matter of time before The Post-Launch Lobotomy. Shame." This sentiment echoes a common theme among users who have observed a decline in their favorite AI tools' effectiveness over time.
Pranjal Paliwal, another developer who had initially praised Astra, revisited the model's generated code and remarked, "We don't have AGI. We have a regression," highlighting his disillusionment. Users have reported issues such as slower response times and diminished quality, with some suggesting that OpenAI may have reduced the model's computational intensity since its launch.
Notably, users have conducted tests comparing Astra's output from launch day to its current performance, with many finding the results lacking. Developer Md Ismail Sojal noted, "The difference is bigger than I expected," after running the same prompt on both versions. Such findings have prompted some users, including Dax Raad, to revert to the previous model, GPT-5.6 Sol, due to dissatisfaction with Astra's performance relative to its costs.
While many users are vocal about their grievances, not all believe that OpenAI intentionally downgraded Astra. Some argue that the initial excitement surrounding the launch led to inflated expectations, and that the model's performance is being judged more critically now that users have had time to evaluate its outputs.
OpenAI has yet to comment on the situation, but the backlash is reminiscent of the criticism faced by the earlier model, GPT-5.6 Sol, which also experienced a similar decline in perceived reasoning capabilities shortly after its release. Despite the controversy, Astra remains priced at $10 per million input tokens and $50 per million output tokens, significantly higher than its predecessor's launch costs. As users continue to express their concerns, many are left wondering about the future performance of this latest AI model.
