Model components / Training & synthetic datasets

Hermes 3 Dataset

Nous Research's release of the instruction and chat corpus behind Hermes 3, mixing synthetic generation with curated community data. It seeds the unreleased Hermes 4 corpus and documents the data recipe behind the Hermes line.

Verified 2026-08-13 via the NousResearch/Hermes-3-Dataset dataset card on Hugging Face.

Openness

5 high confidence

5.0

license: Apache-2.0
gated: false
dataset_card: present

Apache-2.0, ungated, dataset card.

https://huggingface.co/datasets/NousResearch/Hermes-3-Dataset recorded 2026-08-13
Apache-2.0 license, ungated, dataset card

Adoption

1 high confidence

1.0

688 downloads in the trailing 30 days for NousResearch/Hermes-3-Dataset, which bands at <1K, level 1 on the dataset adoption scale.

https://huggingface.co/api/datasets/NousResearch/Hermes-3-Dataset recorded 2026-08-13
688 downloads in the trailing 30 days for NousResearch/Hermes-3-Dataset

Capability

4 high confidence

4.0

Proven value with documented Hermes 3 results and explicitly folded into Hermes 4 (2025), but independent reuse is limited to community quantizers. Table 5 of the Hermes 3 report puts Hermes 3 ahead of Llama 3.1 Instruct on the majority of its seventeen benchmarks at all three sizes.

https://arxiv.org/pdf/2508.18255 recorded 2026-08-13
Hermes 4 report: a significant portion of the Hermes 3 dataset was retained
https://arxiv.org/pdf/2408.11857 recorded 2026-08-13
Hermes 3 beats Llama 3.1 Instruct on most benchmarks at 405B/70B/8B

Verified 2026-08-13