Claude Opus 4.8 Is Nothing More Than Marketing

Opus 4.8 is presented as Anthropic’s new flagship AI model released about a month after Opus 4.7. Anthropic claims it is the most capable model available and provides benchmark results intended to show it outperforms competitors. After using Opus 4.8, quality concerns are raised, stating the model works but still struggles with fundamental issues seen across recent Opus versions. The claim is that Opus quality has declined since Opus 4.6, supported by independent testing. OpenMark testing results attributed to a Reddit user are cited to argue Opus 4.8 performs worse than GPT-5.4 and Opus 4.6.

"Opus 4.8 is the new flagship AI model from Anthropic. This model was released just a month after the previous flagship model, Opus 4.7, and Anthropic claims it is the most capable AI model available on the market."

"To prove this, the Anthropic team shares a nice set of benchmarks showing that Opus 4.8 outperforms its competitors."

"But after spending time working with Opus 4.8, I started to notice that this model is just nothing more than marketing. It works fine, but it still struggles with the fundamental issues that all the latest Opus models have."

"In my experience, Opus model quality has been going downhill since Opus 4.6 and independent tests prove this. Below is a result of the OpenMark testing conducted by one of the Reddit users, and this testing clearly shows that Opus 4.8 is nowhere near the GPT-5.4 and Opus 4.6"

Read at Medium