Skip to main content

AI Bosala

العربية

Head-to-head comparison: Hermes 3 or LongLLaMA?

· 1 min read

In the world of open-source language models, Hermes 3 and LongLLaMA are competing to deliver a more human-like AI experience. The former focuses on logical reasoning and intelligent function invocation, while the latter emphasizes its strength in processing long texts and retaining context over thousands of tokens. Both offer a different look at the future of open models.
Check out the slides below to find out which one is right for you.

  1. Head-to-head comparison: Hermes 3 or LongLLaMA? - Image 1
  2. Head-to-head comparison: Hermes 3 or LongLLaMA? - Image 2
  3. Head-to-head comparison: Hermes 3 or LongLLaMA? - Image 3
  4. Head-to-head comparison: Hermes 3 or LongLLaMA? - Image 4
  5. Head-to-head comparison: Hermes 3 or LongLLaMA? - Image 5
  6. Head-to-head comparison: Hermes 3 or LongLLaMA? - Image 6
  7. Head-to-head comparison: Hermes 3 or LongLLaMA? - Image 7
  8. Head-to-head comparison: Hermes 3 or LongLLaMA? - Image 8
  9. Head-to-head comparison: Hermes 3 or LongLLaMA? - Image 9
  10. Head-to-head comparison: Hermes 3 or LongLLaMA? - Image 10

About Hermes 3

  • Open-weight LLM fine-tuned from Llama 3.1 in 8B, 70B and 405B sizes
  • Long-context retention and multi-turn conversation
  • Complex role-playing
  • Agentic function calling via XML tags
  • Internal monologue and step-wise planning
  • Structured outputs such as JSON and Mermaid diagrams
  • Downloadable from Hugging Face

About LongLLaMA

  • Processes and generates text with contexts up to 256k tokens
  • Built on OpenLLaMA with the Focused Transformer method
  • Retrieves passkeys reliably in very long inputs
  • Includes pre-training and instruction-tuning code
  • Provides inference examples in PyTorch and JAX/Flax
  • Integrates with Hugging Face