Anthropic releases Claude Opus 4.8
anthropic released claude opus 4.8, an incremental update featuring reduced code flaw hallucinations by 4x, mid-conversation system prompts, and a lower cache minimum.
anthropic shipped claude opus 4.8, characterizing the update as a modest but tangible improvement focused on honesty and reducing hallucinations. evaluations indicate the model is four times less likely to let code flaws pass unremarked and achieves the lowest incorrect-rate on benchmarks by abstaining from uncertain answers. technical updates introduce mid-conversation system messages, allowing developers to update instructions in long-running agentic loops without invalidating the prompt cache. additionally, the minimum cacheable prompt length dropped from 4,096 to 1,024 tokens.