Claude Opus 5: the AI leap nobody asked for

Claude Opus 5: the world's most powerful AI model is here, but what's it for?

Anthropic has just launched Claude Opus 5, barely four days after presenting Fable 5. The company is outdoing itself at a pace that puzzles even the most enthusiastic observers. This new model is being promoted as the most powerful ever created, but the noise around these releases raises suspicions.

More power or more marketing?

Token consumption is similar to Opus 4.8 and notably lower than Fable 5, according to early tests. That means one of Claude's weak points —its usage limits— might finally be improving. But skepticism prevails: some point out that Chinese models like Kimi3 already match or surpass Claude Opus 5 with barely a tenth of the computing power. Distillation, the technique that allows large models to train smaller, more efficient ones, is emerging as the real revolution.

The awkward question: real usefulness

Beyond the benchmarks, the discussion centers on practical value. The long-standing complaints about rapid token consumption and the lack of transformative applications —beyond the tech bubble— resurface. Recursive automation, where AI improves itself without human intervention, is mentioned as the true milestone that would change the rules of the game, but it hasn't arrived yet.

Claude Opus 5 is, without a doubt, a technical advance. But public perception remains stubborn: models every four days, staggering consumption figures, and a feeling that the industry is entertaining people with promises while housing, jobs, and daily life don't improve one bit. Are we looking at the engine of the next decade, or the biggest tech smokescreen in history?