Coldbrew GGUF models

137 GGUF models in the Coldbrew family, quantized by 4 uploaders. Sizes run from 40M to 14.8B parameters. The longest trained context is 1000K tokens.

TinySmallMediumLarge
Most downloadedNewestLargestLongest contextSmallest
CO
Q2.5 ColdBrew R1 ForgeWide SAFE
Coldbrew · mradermacher
Params9.5B
Context128K
Q4_K_M5.8 GB
Reasoning
CO
ColdBrew Nemo 12B Arcane Fusion RPThink0
Coldbrew · mradermacher
Params12.2B
Context1000K
Q4_K_M7.5 GB
GGUF
CO
ColdBrew Nemo 12B Arcane Fusion Combined Thinker
Coldbrew · mradermacher
Params12.2B
Context1000K
Q4_K_M7.5 GB
Reasoning
CO
Q2.5 ColdBrew 14B Base RP
Coldbrew · mradermacher
Params14.8B
Context128K
Q4_K_M9.0 GB
RoleplayTool use
CO
MQ Coldbrew Base
Coldbrew · mradermacher
Params7.61B
Context128K
Q4_K_M4.7 GB
GGUF
CO
Q3.5 ColdBrew 9B Base v2 Intrepid
Coldbrew · mradermacher
Params8.95B
Context256K
Q4_K_M5.6 GB
Vision
CO
L3 ColdBrew Daybreak i1
Coldbrew · mradermacher
Params8.03B
Context8K
Q4_K_M4.9 GB
GGUF
CO
Q2.5 ColdBrew RxB
Coldbrew · mradermacher
Params9.24B
Context128K
Q4_K_M5.7 GB
GGUF
CO
Q2.5 ColdBrew R1 Atrax i1
Coldbrew · mradermacher
Params9.24B
Context128K
Q4_K_M5.7 GB
Reasoning
CO
ColdBrew Nemo 12B Arcane Base M
Coldbrew · mradermacher
Params12.2B
Context128K
Q4_K_M7.5 GB
Tool use
CO
Q2.5 ColdBrew RxVm
Coldbrew · mradermacher
Params9.24B
Context128K
Q4_K_M5.7 GB
GGUF
CO
L3 ColdBrew DualCore
Coldbrew · mradermacher
Params8.03B
Context8K
Q4_K_M4.9 GB
GGUF
CO
Q2.5 ColdBrew Antrax Forge 10B
Coldbrew · mradermacher
Params9.94B
Context128K
Q4_K_M6.1 GB
GGUF
CO
ColdBrew Nemo 12B Arcane Fusion KBThink0
Coldbrew · mradermacher
Params12.2B
Context1000K
Q4_K_M7.5 GB
GGUF
CO
Q2.5 ColdBrew R1 Obsidian i1
Coldbrew · mradermacher
Params9.24B
Context128K
Q4_K_M5.7 GB
Reasoning
CO
Q3.5 ColdBrew 9B Base v2 Intrepid
Coldbrew · mradermacher
Params8.95B
Context256K
Q4_K_M5.6 GB
GGUF
CO
Q3.5 ColdBrew 9B Base v2
Coldbrew · mradermacher
Params8.95B
Context256K
Q4_K_M5.6 GB
ReasoningTool use
CO
Q2.5 ColdBrew RxV
Coldbrew · mradermacher
Params9.24B
Context128K
Q4_K_M5.7 GB
GGUF
CO
ColdBrew Nemo 12B Arcane ThinkTest0
Coldbrew · mradermacher
Params12.2B
Context128K
Q4_K_M7.5 GB
Reasoning
CO
Q2.5 ColdBrew 15B Oxford
Coldbrew · mradermacher
Params14.8B
Context128K
Q4_K_M9.0 GB
Tool use
CO
ColdBrew Aphid
Coldbrew · mradermacher
Params6.22B
Context128K
Q4_K_M3.8 GB
Tool use
CO
L3 ColdBrew ReflectRedux
Coldbrew · mradermacher
Params8.03B
Context8K
Q4_K_M4.9 GB
GGUF
CO
Q2.5 Coldbrew14B FusionMix
Coldbrew · mradermacher
Params14.8B
Context128K
Q4_K_M9.0 GB
GGUF
CO
ColdBrew Nemo 12B Arcane Fusion Combined Thinker
Coldbrew · mradermacher
Params12.2B
Context1000K
Q4_K_M7.5 GB
Reasoning
CO
Q3.5 ColdBrew 9B Base v2
Coldbrew · mradermacher
Params8.95B
Context256K
Q4_K_M5.6 GB
ReasoningTool useVision
CO
L3.2ColdBrew ChattyRP i1
Coldbrew · mradermacher
Params3.21B
Context128K
Q4_K_M2.0 GB
GGUF
CO
L3 ColdBrew OldSpice i1
Coldbrew · mradermacher
Params9.78B
Context8K
Q4_K_M6.0 GB
GGUF
CO
Q2.5 ColdBrew R1 Forge_V2
Coldbrew · mradermacher
Params9.48B
Context128K
Q4_K_M5.8 GB
Reasoning
CO
ColdBrew Nemo 12B Arcane Fusion KBThink0
Coldbrew · mradermacher
Params12.2B
Context1000K
Q4_K_M7.5 GB
GGUF
CO
Q2.5 ColdBrew R1 Indigo i1
Coldbrew · mradermacher
Params9.24B
Context128K
Q4_K_M5.7 GB
Reasoning

About the Coldbrew GGUF models

Coldbrew is represented on ggufmodels.com by 137 GGUF models, quantized by 4 uploaders on Hugging Face. The most downloaded are Q2.5 ColdBrew R1 ForgeWide SAFE, ColdBrew Nemo 12B Arcane Fusion RPThink0 and ColdBrew Nemo 12B Arcane Fusion Combined Thinker. Parameter counts run from 40M to 14.8B, and the longest trained context window is 1000K tokens.

A Q4_K_M file of a Coldbrew model is between 2.0 to 9.0 GB, which is the quant most people start with. Most files come from mradermacher (78), Theros (45) and SvalTek (13). The quants you will meet most often are Q4_K_M (131), Q5_K_M (94), Q6_K (79) and Q5_K_S (79), where the number is how many models offer that quant.

Frequently asked questions

Which Coldbrew GGUF should I download?

The most downloaded Coldbrew GGUF here is Q2.5 ColdBrew R1 ForgeWide SAFE. Pick the size that fits your GPU first, then choose a quant: Q4_K_M is the usual starting point, and Q5_K_M or Q6_K give more quality if you have memory to spare. The VRAM pages below list what fits each GPU.

How much VRAM does a Coldbrew model need?

Q4_K_M files of Coldbrew models on this site range from 2.0 to 9.0 GB. Add the KV cache for your context length and about 0.5 GB for the runtime, which usually means 1.5 to 3 GB on top of the file at an 8K context.

Where do Coldbrew GGUF files come from?

The most common quantizers here are mradermacher, Theros and SvalTek. They convert the original model and publish quantized files on Hugging Face. ggufmodels.com does not host files; every download button opens Hugging Face.

Keep browsing