Claude’s new model is more ‘honest’ when it messes up

Anthropic released Claude Opus 4.8, an AI model that’s designed to be more honest about its limitations. The company trains its models to avoid making unsupported claims, but acknowledges that AI sometimes jumps to conclusions. This new version aims to be clearer about when it’s unsure.

This matters because many people get frustrated when AI seems confident but is actually wrong. Imagine asking your phone for directions and it confidently sends you the wrong way—annoying, right? With this update, Claude is more likely to say, “I’m not sure,” which could make AI feel more reliable.

If you use Claude, try asking it a tricky question today. For example, ask, “What’s the capital of a fictional country?” and see how it responds. You can find Claude at claude.ai.