Hermes: Hermes Agent Model

Use as the model behind Nous Research's Hermes Agent (custom OpenAI-compatible endpoint). Send Hermes's own system prompt and tools unchanged; Neurometric Hermes adds guidance on top so the agent finishes multi-step work with tools, asks before destructive commands and production actions, runs the tests after editing code and never claims success on a red suite, follows the project's AGENTS.md rules, and answers plainly for the CLI.

Compare measured quality, workload cost, and latency across model tiers. Results are task-specific and directional—not a general model ranking.