Anthropic has released Claude Opus 4.8, an updated version of its artificial intelligence that maintains the previous price. The model stands out for pointing out errors in code with greater transparency and reduces usage costs by up to 61% in specific tasks. For those who use AI in their daily work, this translates into more reliable and faster responses without spending more money. A practical improvement that brings artificial intelligence closer to real needs.
Fewer hallucinations, more savings per query 💰
The new version of Opus incorporates a more critical approach when analyzing code, recognizing its own limitations and pointing out potential flaws instead of giving false answers. This reduces the need for subsequent human reviews. Additionally, optimization in processing allows lowering the cost per task by up to 61% in cases such as script debugging or report generation. An update that prioritizes usefulness over grandiosity, offering more efficient performance for developers and technical users seeking accurate results without paying extra.
The AI that admits its mistakes (and even charges you less) 😅
Finally, an artificial intelligence that doesn't think it's perfect. Claude Opus 4.8 tells you this is wrong and even charges you less for the privilege. A novelty in a sector where models usually respond with confidence even when they are making up data. Now, when the code fails, at least you'll have someone (or something) that confesses it to you without beating around the bush and at a discount. It almost seems like a digital therapist: it tells you the truth and doesn't bankrupt you in the process.