Hermes 3 405B

The Hermes 3 flagship: a 405B fine-tuneA model that has been further trained on additional data to specialize it for a particular task, domain, or style. Fine-tuning a general model on medical literature produces a medical specialist; fine-tuning on your company's support tickets produces a support assistant that sounds like your team. Fine-tunes are much cheaper to create than training a model from scratch. of Llama 3.1 405B that was state-of-the-art among open-weightA model where the trained weights are freely downloadable — you can run it yourself without contacting the creator. Llama, Mistral, Qwen, and Gemma are open-weight. Open-weight does not mean open-source: the training data and code often stay private. The license still governs what you can do with the weights, including whether you can use them commercially. models at its 2024 release — now succeeded by Hermes 4.

Hermes 3 405B was Nous Research's headline model before Hermes 4, and at its August 2024 launch it posted state-of-the-art results among open-weightA model where the trained weights are freely downloadable — you can run it yourself without contacting the creator. Llama, Mistral, Qwen, and Gemma are open-weight. Open-weight does not mean open-source: the training data and code often stay private. The license still governs what you can do with the weights, including whether you can use them commercially. models on several public benchmarks. It introduced the agentic and structured-output capabilities — XML, strong multi-turn coherence, long-context retention — that define the modern Hermes line.

It remains a capable model, but Hermes 4 405B supersedes it with hybrid reasoning and sharper outputs, so for new work the 4 line is the better starting point. This entry mainly documents the lineage and supports teams already running it.

License is inherited from Llama (see below).

• Existing deployments standardized on Hermes 3 405B.
• Understanding the Hermes lineage and the jump to Hermes 4.
• Agentic/structured-output work at frontier scale where a 3-line model is already validated.

• New deployments — start on Hermes 4 405B.
• Anyone without cluster-scale hardware — use a smaller Hermes.
• Products near 700M MAU (Llama carve-out).
• MultimodalA model that can handle more than one type of input or output — typically text plus images, sometimes plus audio or video. "GPT-4 Vision" and "Llama 3.2 11B Vision" are multimodal models that accept both text and images. A text-only model is called "unimodal" but nobody uses that term; text-only is the assumed default. tasks — text only.

Two layers. Nous's open Hermes 3 weightsThe numerical values inside a trained model that encode everything it has learned. A model is, functionally, a giant list of weights — tens of billions of numbers for a mid-sized model, hundreds of billions for a frontier model. "Open-weight" means those numbers are published. "Downloading the weights" means getting the actual file you'd need to run the model yourself. sit on Meta's Llama 3.1 405B, so the Llama 3.1 Community License governs — commercial use allowed with "Built with Llama" attribution and the 700M-MAU carve-out. Same terms as the rest of the Llama-based Hermes family.

Against Hermes 4 405B, Hermes 3 405B is the older flagship — Hermes 4 is the upgrade. Against Hermes 3 70B, it's the larger, heavier sibling of the same generation. Against its base Llama 3.1 405B, it's Nous's steerable tuning rather than Meta's Instruct.