Hermes 3 Technical Report

TL;DR

Hermes 3 is a neutrally-aligned, generalist instruct and tool-use model with strong reasoning and creative abilities, achieving state-of-the-art performance among open weight models on various benchmarks.

Contribution

Introduces Hermes 3, a new instruct-tuned model with strong reasoning and creative skills, and demonstrates its superior performance on multiple benchmarks.

Findings

Hermes 3 405B achieves state-of-the-art results among open weight models.

The model demonstrates strong reasoning capabilities.

It is neutrally aligned and versatile in instruction and tool use.

Abstract

Instruct (or "chat") tuned models have become the primary way in which most people interact with large language models. As opposed to "base" or "foundation" models, instruct-tuned models are optimized to respond to imperative statements. We present Hermes 3, a neutrally-aligned generalist instruct and tool use model with strong reasoning and creative abilities. Its largest version, Hermes 3 405B, achieves state of the art performance among open weight models on several public benchmarks.