Every one of the distinct chat-capable model IDs currently served through Nous Portal's inference API — pulled live from its own /v1/models endpoint, not Hermes's curated picker shortlist. carry a published Artificial Analysis Intelligence Index reading; the rest are ranked by real spec data only (price, context, tool-calling, reasoning, modality, architecture) — never a guessed score. Pulled .
Price vs. intelligence
the models with a published score · output price, log scale · bubble = context window · hover any point