Hermes can absolutely be S-tier. But only if you ask: S-tier for WHAT? 🪽 If the job is pure coding, compare coding specialists. If it’s deep research, compare research systems. But if you want an agent that can: → keep working on your PC/server → use tools and browse → run scheduled jobs → handle multi-step workflows → stay available beyond one chat session Hermes starts looking very different. That’s why one giant AI-agent tier list misses the point. Hermes isn’t trying to win every category. It gets interesting when you want a real operator, not just a single-purpose assistant. Pick the job first. Then pick the agent. For persistent agent work, Hermes belongs near the top of the conversation. Which category should we test Hermes against next?