LL

Hermes 3 Llama 3.1 405B

LlamaGGUF by bartowskiUpdated 2024-09-27
Download on Hugging Face
Parameters405B
Context128K
Layers126
Architecturellama
Smallest quant93.5 GB
Quick answer

Hermes 3 Llama 3.1 405B needs about 247.8 GB of memory at Q4_K_M (a 243.1 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 20 quants, from 93.5 GB (IQ1_M) to 431.2 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 16.9 GB.

Hermes 3 Llama 3.1 405B – a large generalist language model

Hermes 3 Llama 3.1 405B is a 405‑billion‑parameter model fine‑tuned from Llama‑3.1 405B. It supports roleplay, tools, and long‑context conversation up to 131072 tokens. The model is released under the llama3 license.

Good at

  • Long‑context reasoning
  • Agentic tool use
  • Structured output

Limits

  • No vision support
  • Requires large VRAM for full quality
  • License is llama3

Best for

  • Complex dialogue systems
  • AI‑powered assistants
  • Code generation tasks

Who it’s for. Developers building advanced conversational agents or assistants who need a large, generalist model with strong reasoning and tool‑calling abilities. The model is suitable for users with high‑end GPUs or sufficient RAM to run the larger quantizations.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ1_M93.5 GB (3 parts) 98.2 GB
Too big Get file →
IQ1_M93.5 GB (10 parts) 98.2 GB
Too big Get file →
IQ2_XXS107.3 GB (3 parts) 112.0 GB
Too big Get file →
IQ2_XS119.3 GB (3 parts) 124.1 GB
Too big Get file →
IQ2_M136.7 GB (4 parts) 141.4 GB
Too big Get file →
Q2_K149.3 GB (4 parts) 154.1 GB
Too big Get file →
Q2_K_L151.4 GB (4 parts) 156.1 GB
Too big Get file →
IQ3_XXS155.8 GB (4 parts) 160.6 GB
Too big Get file →
Q3_K_S175.2 GB (5 parts) 180.0 GB
Too big Get file →
IQ3_M181.7 GB (5 parts) 186.5 GB
Too big Get file →
Q3_K_M195.4 GB (5 parts) 200.1 GB
Too big Get file →
Q3_K_L212.8 GB (6 parts) 217.6 GB
Too big Get file →
Q3_K_XL214.7 GB (6 parts) 219.4 GB
Too big Get file →
IQ4_XS216.6 GB (6 parts) 221.3 GB
Too big Get file →
Q4_K_S230.5 GB (6 parts) 235.2 GB
Too big Get file →
Q4_K_M243.1 GB (7 parts) 247.8 GB
Too big Get file →
Q4_K_L244.6 GB (7 parts) 249.4 GB
Too big Get file →
Q5_K_M286.6 GB (8 parts) 291.4 GB
Too big Get file →
Q6_K332.9 GB (9 parts) 337.7 GB
Too big Get file →
Q8_0431.2 GB (11 parts) 436.0 GB
Too big Get file →

For 8‑GB VRAM, use Q4_K_S (230.5 GB file, 235.2 GB memory at 8k context). For 16‑GB VRAM, Q4_K_M (243.07 GB file, 247.8 GB memory) is a good balance. For 24‑GB or more, Q5_K_M (286.65 GB file, 291.4 GB memory) offers higher quality. Memory estimates are approximate.

What GPU do you need for Hermes 3 Llama 3.1 405B?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: no quant fits at an 8K context.
  • 6 GB VRAM: no quant fits at an 8K context.
  • 8 GB VRAM: no quant fits at an 8K context.
  • 12 GB VRAM: no quant fits at an 8K context.
  • 16 GB VRAM: no quant fits at an 8K context.
  • 24 GB VRAM: no quant fits at an 8K context.
  • 32 GB VRAM: no quant fits at an 8K context.
  • 48 GB VRAM: no quant fits at an 8K context.

Which Hermes 3 Llama 3.1 405B quant should you choose?

Q4_K_M is the usual starting point for Hermes 3 Llama 3.1 405B: it balances file size and quality well. Here it is a 243.1 GB file.

The Q8_0 file is 431.2 GB, about 1.8 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.

If memory is tight, the smallest file is IQ1_M at 93.5 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturellama
Parameters405B
Layers126
Trained context131,072 tokens
KV cache (fp16)504 KB per token

Source

Quantized bybartowski
FamilyLlama
Licensellama3
Downloads (HF)3,518
Last updated2024-09-27

Frequently asked questions about Hermes 3 Llama 3.1 405B

How much VRAM does Hermes 3 Llama 3.1 405B need?

About 247.8 GB at Q4_K_M with an 8K context window. That is the 243.1 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run Hermes 3 Llama 3.1 405B on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 98.2 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Hermes 3 Llama 3.1 405B GGUF?

It depends on the quant: IQ1_M is 93.5 GB, IQ1_M is 93.5 GB, IQ2_XXS is 107.3 GB, IQ2_XS is 119.3 GB, IQ2_M is 136.7 GB, Q2_K is 149.3 GB and 14 more.

What context length does Hermes 3 Llama 3.1 405B support?

The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 4.2 GB at 8K and 16.9 GB at 32K.

Where can I download Hermes 3 Llama 3.1 405B GGUF?

The files are on Hugging Face in the bartowski/Hermes-3-Llama-3.1-405B-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does Hermes 3 Llama 3.1 405B use?

Hugging Face lists the license as llama3. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run Hermes 3 on 8 GB VRAM?

Use the Q4_K_S quant, which requires about 235.2 GB of memory at 8k context. It will fit on 8 GB VRAM with sharding.

Does Hermes 3 support vision?

No, the model does not include vision capabilities.

What license applies to Hermes 3?

The model is released under the llama3 license.

Other GGUFs of Hermes-3-Llama-3.1-405B

Similar models

Browse all →

More from bartowski

All by bartowski →

More Llama models

All Llama →

Other large models

All large models →