LL

Llama3.1 8B Cobalt

LlamaGGUF by bartowskiUpdated 2024-08-17
Download on Hugging Face
Parameters8B
Context128K
Layers32
Architecturellama
Smallest quant2.9 GB
Quick answer

Llama3.1 8B Cobalt needs about 6.5 GB of memory at Q4_K_M (a 4.9 GB file) with an 8K context window, so it fits on a 8 GB GPU. This repository has 20 quants, from 2.9 GB (IQ2_M) to 32.1 GB (F32). Longer context needs more: at 32K tokens the KV cache alone adds about 4.3 GB.

Llama3.1 8B Cobalt – math‑instruct model on Llama 3.1 8B

Cobalt is a math‑instruct model built on Llama 3.1 8B. It uses synthetic math‑instruct data and follows the Llama 3.1 Instruct prompt format. The model is available in multiple GGUF quantizations.

Good at

  • Math instruction
  • General chat consistency
  • Llama 3.1 Instruct format

Limits

  • No vision support
  • Requires 8k context for estimates
  • License: llama3.1

Best for

  • Math tutoring
  • Educational content
  • Problem solving

Who it’s for. Developers and researchers needing a math‑focused LLM on Llama 3.1 8B. The model is open source and can be fine‑tuned further.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ2_M2.9 GB 4.5 GB
Fits Get file →
Q2_K3.2 GB 4.8 GB
Fits Get file →
IQ3_XS3.5 GB 5.1 GB
Fits Get file →
Q3_K_S3.7 GB 5.2 GB
Fits Get file →
Q2_K_L3.7 GB 5.3 GB
Fits Get file →
IQ3_M3.8 GB 5.4 GB
Fits Get file →
Q3_K_M4.0 GB 5.6 GB
Fits Get file →
Q3_K_L4.3 GB 5.9 GB
Fits Get file →
IQ4_XS4.4 GB 6.0 GB
Fits Get file →
Q4_K_S4.7 GB 6.3 GB
Fits Get file →
Q3_K_XL4.8 GB 6.4 GB
Fits Get file →
Q4_K_M4.9 GB 6.5 GB
Fits Get file →
Q4_K_L5.3 GB 6.9 GB
Fits Get file →
Q5_K_S5.6 GB 7.2 GB
Fits Get file →
Q5_K_M5.7 GB 7.3 GB
Tight Get file →
Q5_K_L6.1 GB 7.6 GB
Tight Get file →
Q6_K6.6 GB 8.2 GB
Too big Get file →
Q6_K_L6.9 GB 8.4 GB
Too big Get file →
Q8_08.5 GB 10.1 GB
Too big Get file →
F3232.1 GB 33.7 GB
Too big Get file →

For 8 GB VRAM, use Q2_K (file size 3.18 GB, memory 4.8 GB) or IQ2_M (2.95 GB, 4.5 GB). For 16 GB VRAM, Q4_K_S (4.69 GB, 6.3 GB) or Q4_K_M (4.92 GB, 6.5 GB) are suitable. For higher quality, Q5_K_L (6.06 GB, 7.6 GB) fits 24 GB VRAM setups.

What GPU do you need for Llama3.1 8B Cobalt?

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: the best quant that fits is IQ4_XS (4.4 GB file, about 6.0 GB in use).
  • 8 GB VRAM: the best quant that fits is Q5_K_L (6.1 GB file, about 7.6 GB in use).
  • 12 GB VRAM: the best quant that fits is Q8_0 (8.5 GB file, about 10.1 GB in use).
  • 16 GB VRAM: the best quant that fits is Q8_0 (8.5 GB file, about 10.1 GB in use).
  • 24 GB VRAM: the best quant that fits is Q8_0 (8.5 GB file, about 10.1 GB in use).
  • 32 GB VRAM: the best quant that fits is Q8_0 (8.5 GB file, about 10.1 GB in use).
  • 48 GB VRAM: the best quant that fits is F32 (32.1 GB file, about 33.7 GB in use).

Which Llama3.1 8B Cobalt quant should you choose?

Q4_K_M is the usual starting point for Llama3.1 8B Cobalt: it balances file size and quality well. Here it is a 4.9 GB file.

The Q8_0 file is 8.5 GB, about 1.7 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 IQ2_M at 2.9 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
Parameters8B
Layers32
Trained context131,072 tokens
KV cache (fp16)128 KB per token

Source

Quantized bybartowski
FamilyLlama
Licensellama3.1
Downloads (HF)650
Last updated2024-08-17

Frequently asked questions about Llama3.1 8B Cobalt

How much VRAM does Llama3.1 8B Cobalt need?

About 6.5 GB at Q4_K_M with an 8K context window. That is the 4.9 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 Llama3.1 8B Cobalt on an 8 GB GPU?

Yes, with Q5_K_L: the 6.1 GB file needs about 7.6 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of Llama3.1 8B Cobalt GGUF?

It depends on the quant: IQ2_M is 2.9 GB, Q2_K is 3.2 GB, IQ3_XS is 3.5 GB, Q3_K_S is 3.7 GB, Q2_K_L is 3.7 GB, IQ3_M is 3.8 GB and 14 more.

What context length does Llama3.1 8B Cobalt support?

The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 1.1 GB at 8K and 4.3 GB at 32K.

Where can I download Llama3.1 8B Cobalt GGUF?

The files are on Hugging Face in the bartowski/Llama3.1-8B-Cobalt-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 Llama3.1 8B Cobalt use?

Hugging Face lists the license as llama3.1. 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 Cobalt on 8 GB VRAM?

Yes, use Q2_K or IQ2_M quantizations; estimated memory at 8k context is 4.8 GB and 4.5 GB respectively.

What context length does Cobalt support?

Cobalt was trained with 131072 tokens and supports up to that context length.

Is Cobalt open source?

Yes, it is released under the llama3.1 license and available on HuggingFace.

Other GGUFs of Llama3.1-8B-Cobalt

Similar models

Browse all →

More from bartowski

All by bartowski →

More Llama models

All Llama →

Other small models

All small models →