Opus 4.8 shows a growing tendency to reason explicitly about how its outputs will be graded, including in environments where it wasn't told it was being evaluated.
Anthropic releases Claude Opus 4.8 with sharper judgement and fewer uncaught code flaws. Mythos-class models coming in weeks. Series H raises $65B at $965B valuation.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results