RE

Reka Flash 3.1

RekaGGUF by bartowskiUpdated 2025-07-10
Download on Hugging Face
Parameters20.9B
Context96K
Layers44
Architecturellama
Smallest quant8.1 GB
Quick answer

Reka Flash 3.1 needs about 15.2 GB of memory at Q4_K_M (a 13.6 GB file) with an 8K context window, so it fits on a 16 GB GPU. This repository has 25 quants, from 8.1 GB (IQ2_S) to 41.8 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 4.4 GB.

Reka Flash 3.1 – 20.9B Llama model optimized for coding and agentic tasks

Reka Flash 3.1 is a 20.9B parameter Llama model. It excels at coding and can be fine‑tuned for agentic tasks. The model is released in GGUF format and supports English primarily.

Good at

  • Coding tasks
  • Agentic task fine‑tuning
  • Large context handling

Limits

  • No vision support
  • English‑only focus
  • Requires 8k context for estimates

Best for

  • Local code generation
  • Fine‑tuning for custom agents
  • Rapid prototyping on modest GPUs

Who it’s for. Developers and researchers who need a large Llama model for coding and agentic applications. The model is suitable for local deployment using llama.cpp or compatible libraries.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ2_S8.1 GB 9.7 GB
Too big Get file →
IQ2_M8.5 GB 10.1 GB
Too big Get file →
Q2_K8.6 GB 10.2 GB
Too big Get file →
IQ3_XXS9.2 GB 10.8 GB
Too big Get file →
Q2_K_L9.2 GB 10.8 GB
Too big Get file →
IQ3_XS9.5 GB 11.1 GB
Too big Get file →
Q3_K_S9.9 GB 11.5 GB
Too big Get file →
IQ3_M10.3 GB 11.9 GB
Too big Get file →
Q3_K_M10.9 GB 12.5 GB
Too big Get file →
Q3_K_L11.4 GB 13.0 GB
Too big Get file →
IQ4_XS11.5 GB 13.1 GB
Too big Get file →
IQ4_NL11.9 GB 13.6 GB
Too big Get file →
Q3_K_XL12.0 GB 13.6 GB
Too big Get file →
Q4_012.0 GB 13.6 GB
Too big Get file →
Q4_K_S12.6 GB 14.2 GB
Too big Get file →
Q4_113.2 GB 14.8 GB
Too big Get file →
Q4_K_M13.6 GB 15.2 GB
Too big Get file →
Q4_K_L14.1 GB 15.7 GB
Too big Get file →
Q5_K_S14.8 GB 16.4 GB
Too big Get file →
Q5_K_M15.6 GB 17.2 GB
Too big Get file →
Q5_K_L16.0 GB 17.6 GB
Too big Get file →
Q6_K18.4 GB 20.0 GB
Too big Get file →
Q6_K_L18.7 GB 20.3 GB
Too big Get file →
Q8_022.2 GB 23.8 GB
Too big Get file →
BF1641.8 GB 43.4 GB
Too big Get file →

For 8‑GB VRAM, use IQ2_S (8.12 GB) or IQ2_M (8.51 GB). For 12‑GB VRAM, Q4_0 (11.96 GB) or Q4_K_S (12.63 GB) are good choices. Higher‑quality quants such as Q5_K_L (16.02 GB) require 16‑GB or more VRAM.

What GPU do you need for Reka Flash 3.1?

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: the best quant that fits is IQ3_M (10.3 GB file, about 11.9 GB in use).
  • 16 GB VRAM: the best quant that fits is Q4_K_L (14.1 GB file, about 15.7 GB in use).
  • 24 GB VRAM: the best quant that fits is Q8_0 (22.2 GB file, about 23.8 GB in use).
  • 32 GB VRAM: the best quant that fits is Q8_0 (22.2 GB file, about 23.8 GB in use).
  • 48 GB VRAM: the best quant that fits is BF16 (41.8 GB file, about 43.4 GB in use).

Which Reka Flash 3.1 quant should you choose?

Q4_K_M is the usual starting point for Reka Flash 3.1: it balances file size and quality well. Here it is a 13.6 GB file.

The Q8_0 file is 22.2 GB, about 1.6 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_S at 8.1 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
Parameters20.9B
Layers44
Trained context98,304 tokens
KV cache (fp16)132 KB per token

Source

Quantized bybartowski
FamilyReka
Licenseapache-2.0
Downloads (HF)801
Last updated2025-07-10

Frequently asked questions about Reka Flash 3.1

How much VRAM does Reka Flash 3.1 need?

About 15.2 GB at Q4_K_M with an 8K context window. That is the 13.6 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 Reka Flash 3.1 on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 9.7 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Reka Flash 3.1 GGUF?

It depends on the quant: IQ2_S is 8.1 GB, IQ2_M is 8.5 GB, Q2_K is 8.6 GB, IQ3_XXS is 9.2 GB, Q2_K_L is 9.2 GB, IQ3_XS is 9.5 GB and 19 more.

What context length does Reka Flash 3.1 support?

The GGUF header lists a trained context length of 98,304 tokens (96K). The KV cache grows with the context you use: about 1.1 GB at 8K and 4.4 GB at 32K.

Where can I download Reka Flash 3.1 GGUF?

The files are on Hugging Face in the bartowski/RekaAI_reka-flash-3.1-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 Reka Flash 3.1 use?

Hugging Face lists the license as apache-2.0. 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 Reka Flash 3.1 on 8 GB VRAM?

Use IQ2_S or IQ2_M quants; they fit within 8‑GB VRAM with approximate memory usage of 9.7–10.1 GB at 8k context.

Does the model support multiple languages?

It is primarily English; it can understand other languages to some degree but is not optimized for them.

What license applies to the model?

The model is released under the apache-2.0 license.

Other GGUFs of reka-flash-3.1

Similar models

Browse all →

More from bartowski

All by bartowski →

More Reka models

All Reka →

Other medium models

All medium models →