BF16 GGUF models

Unquantized bfloat16. Original quality and the largest size. 9004 models have a BF16 file. Sizes run from 1M to 2780B parameters. The longest trained context is 10240K tokens.

Most downloadedNewestLargestLongest contextSmallest
QW
Qwen2.5-VL-32B-Instruct
Qwen · Mungert
Params32.8B
Context125K
Q4_K_M20.1 GB
Vision
QW
Qwen3.8 27B DSpark
Qwen · DimInfer
Params1.86B
Context256K
Smallest2.0 GB
GGUF
IN
Intern Decision 4B
InternLM · prithivMLmods
Params4.21B
Context256K
Q4_K_M2.7 GB
Vision
GE
Gemma 4 26B A4B It Heretic
Gemma · SassyDiffusion
Params25.2B
Context256K
Q4_K_M16.9 GB
Vision
FA
Fara 7B
Fara · prithivMLmods
Params7.62B
Context125K
Q4_K_M4.7 GB
Vision
QW
Qwen3.5 4B Base ZitGen V1
Qwen · lolzinventor
Params4.21B
Context256K
Smallest4.5 GB
ReasoningTool useVision
QW
tvall43/Qwen3.5-4B-heretic
Qwen · tvall43
Params4.21B
Context256K
Q4_K_M2.7 GB
Vision
QW
Qwen3 VL 4B Thinking
Qwen · bartowski
Params4B
Context256K
Q4_K_M2.5 GB
ReasoningVision
NE
AceReason Nemotron 7B
Nemotron · bartowski
Params7.62B
Context128K
Q4_K_M4.7 GB
CodingReasoningTool use
GR
Granite Embedding 97m Multilingual R2
Granite · mykor
Params97M
Context32K
Q4_K_M0.1 GB
Embedding
PH
Phi 3 medium 128k instruct imatrix
Phi · hermes42
Params14B
Context128K
Q4_K_M8.6 GB
CodingTool use
DE
Deepseek-R1-0528-Qwen3-8B
DeepSeek · cecloud33
Params8.19B
Context128K
Q4_K_M5.0 GB
ReasoningTool use
ZE
zeta2
Zeta · bluevoid-pl
Params8.25B
Context32K
Q4_K_M5.1 GB
GGUF
QW
Logics SWE Qwen3.6 35B A3B
Qwen · prithivMLmods
Params34.7B
Context256K
Smallest18.1 GB
CodingReasoningTool use
SA
Salience 27B R6
Salience · bartowski
Params27.3B
Context1024K
Q4_K_M17.4 GB
CodingReasoningTool use
EX
EXAONE-Deep-7.8B
Exaone · LGAI-EXAONE
Params7.8B
Context32K
Q4_K_M4.8 GB
GGUF
LA
Laguna S 2.1 Uncensored Heretic
Laguna · llmfan46
Params118B
Context1024K
Q4_K_M73.0 GB
Vision
BA
Bathory X 26B A4B 0.2
Bathory · Nabbers1999
Params26B
Context256K
Q4_K_M17.2 GB
ReasoningTool use
MI
MiniCPM4.1-8B
MiniCPM · Mungert
Params8.19B
Context64K
Q4_K_M5.1 GB
Reasoning
LO
Lotus 1
Lotus · yashsutorichat
Params34.7B
Context256K
Q4_K_M21.2 GB
ReasoningRoleplayTool use
TU
turkish pii detection
Turkish · halilneed
Params268M
Context32K
Smallest0.3 GB
GGUF
G9
G9v3 39A5B
linuxid10t
Params39B
Context128K
Q4_K_M23.6 GB
ReasoningTool use
SQ
squad-0.8b documents expert (LoRA r64: lora_A Q2_K = its 2-bit rows, lora_B Q8_0)
Squad · labqoat
Params11M
Context—
Smallest0.0 GB
GGUF
NE
OpenReasoning Nemotron 1.5B
Nemotron · bartowski
Params1.54B
Context128K
Q4_K_M1.0 GB
Reasoning
LU
Lumimaid v0.2 8B
Lumimaid · Lewdiculous
Params8B
Context128K
Q4_K_M4.9 GB
Roleplay
CL
Clover 35B A3B Preview BF16
Clover · IsValorum
Params34.7B
Context128K
Smallest69.4 GB
ReasoningTool use
SA
Sage A1.5
Sage · prithivMLmods
Params35.5B
Context256K
Smallest18.6 GB
CodingReasoningTool use
KA
KAT Dev
Kat · bartowski
Params32.8B
Context128K
Q4_K_M19.8 GB
GGUF
QW
Qwen3 TTS 12Hz 1.7B CustomVoice
Qwen · BricksDisplay
Params694M
Context32K
Q4_K_M1.1 GB
GGUF
SY
Synthia S1 27b
Synthia · bartowski
Params27B
Context128K
Q4_K_M16.5 GB
Vision

Other quants

Keep browsing