Hermes 4.3 36B logo

MODELS

Hermes 4.3 36B

by Nous Research

llmnous-research
Visit WebsiteDocumentationGitHub

Overview

Open-weight 36B hybrid-reasoning language model from Nous Research with long-context support and local-deployment-oriented availability.

Details

Hermes 4.3 36B is an open-weight 36B hybrid reasoning LLM from Nous Research. Official source metadata describes it as supporting long context, structured outputs, and GGUF variants. Nous Portal lists Hermes-4.3-36B as a 128K-context hybrid-thinking model optimized for local deployment, but also lists it under unsupported models, says it is currently unavailable through the Nous API, and shows pricing as N/A. Microsoft Foundry’s catalog identifies the model as vLLM-powered with OpenAI-compatible Chat Completions and recommends the Responses API for this reasoning model.

When to Use

Use when evaluating an open-weight 36B reasoning model for local deployment or offline experimentation. Use when you need a long-context hybrid-thinking model and can work from Hugging Face GGUF variants or a third-party catalog rather than the Nous API. Use when testing OpenAI-compatible Chat Completions behavior for the Microsoft Foundry-hosted model variant.

Getting Started

  1. Read the Nous Research launch page for the Hermes 4.3 36B overview: https://nousresearch.com/introducing-hermes-4-3
  2. Review the Hugging Face model page for model files and usage information: https://huggingface.co/NousResearch/Hermes-4.3-36B
  3. Check Nous Portal model availability before planning API use
  4. because the supplied source says the model is currently unavailable via the Nous API and pricing is N/A.
  5. If using Microsoft Foundry
  6. consult the catalog entry for its vLLM-powered
  7. OpenAI-compatible Chat Completions behavior and Responses API recommendation.

Key Features

  • •36B-parameter open-weight Hermes 4.3 model from Nous Research
  • •Hybrid reasoning / hybrid-thinking model description in official source metadata
  • •128K context listed by Nous Portal
  • •Structured output support described in supplied official source metadata
  • •GGUF variants described in supplied official source metadata
  • •Microsoft Foundry catalog entry identifies a vLLM-powered deployment with OpenAI-compatible Chat Completions

Capabilities

  • •text generation
  • •reasoning
  • •long-context processing
  • •structured outputs
  • •OpenAI-compatible chat completions
  • •local deployment

Last updated Jun 22, 2026