What is Opus 5?
Opus 5 is a frontier AI model from Anthropic, referenced in discussions about model capabilities and usage. As of mid-2026, it is noted for its persistence and robustness, with reports of it running for extended periods and resisting prompt injection.
Release history
- Jul 2026 - Boris Cherny said Opus 5 “can go for days, weeks, months at a time. It just won’t stop.”
- Jul 2026 - Boris Cherny noted that “the model does not seem to be prompt injectable.”
- Aug 2026 - Nathaniel Whittemore reported that a competing model is now the top used closed model on the platform, seeing 40% more use than Opus 5 and Sonnet 5 combined.
In the discourse
Attributed discussion of Opus 5.
Boris Cherny on Opus 5 running indefinitely without stopping.
“It can go for days, weeks, months at a time. It just won't stop.”Boris Cherny · 27 Jul 2026
Luna is the top closed model on OpenRouter with 40% more usage than Opus 5 and Sonnet 5 combined.
“It's now the top used closed model on the platform seeing 40% more use than Opus 5 and Sonnet 5 combined.”Nathaniel Whittemore · 21 Aug 2026
Luna: new closed model reaching top usage on OpenRouter, surpassing Anthropic's Opus 5 and Sonnet 5 combined, making it a model to track for enterprise adoption signals.
“It's now the top used closed model on the platform seeing 40% more use than Opus 5 and Sonnet 5 combined.”Nathaniel Whittemore · 21 Aug 2026
Opus 5 does not appear to be prompt injectable, contradicting the widespread assumption that large language models are inherently vulnerable to prompt injection.
“The model does not seem to be prompt injectable.”Boris Cherny · 27 Jul 2026