Hermes 3 8B

Good fit for

People configuring local multi-turn character conversations or structured-output workflows.

Why consider it

The official model card explicitly discusses roleplaying, steerable system prompts, code generation and tool-call formats.

Main limitations

Function calling requires software that parses and executes calls; the weights do not provide web search or code execution by themselves. Example prompts claiming consciousness are not factual descriptions. Review the exact downloaded variant and license before redistribution.

Cost and access

No subscription or hosted token fee for local inference within permitted uses. Hardware, electricity and setup are separate. The runtime and model card carry Llama community licensing. Follow the included weights license and upstream requirements rather than assuming MIT or Apache terms.

Getting started

Install a compatible Ollama version, review the model license, then run: ollama run hermes3:8b. API means the local runtime API, not a hosted endpoint included with the model.