🔮All Our Models

Unsloth model catalog for all our Dynamic GGUF, 4-bit, 16-bit models on Hugging Face.

• GGUF + 4-bit

DeepSeekLlamaGemmaQwenMistralPhi

GGUFs let you run models in tools like Ollama, Open WebUI, and llama.cpp. Instruct (4-bit) safetensors can be used for inference or fine-tuning.

New & recommended models:

Model	Variant	GGUF	Instruct (4-bit)
Gemma 3n	E2B	link	link
	E4B	link	link
DeepSeek-R1-0528	R1-0528-Qwen3-8B	link	link
	R1-0528	link	—
Mistral	Small 3.2 24B (2506)	link	link
	Magistral Small (2506)	link	link
FLUX.1	Kontext-dev	link	—
Qwen3	0.6 B	link	link
	1.7 B	link	link
	4 B	link	link
	8 B	link	link
	14 B	link	link
	30B-A3B	link	link
	32 B	link	link
	235B-A22B	link	—
Llama 4	Scout 17B 16E	link	link
	Maverick 17B 128E	link	—
Qwen-2.5 Omni	3 B	link	—
	7 B	link	—
Phi-4	Reasoning-plus	link	link
	Reasoning	link	link

DeepSeek models:

Model	Variant	GGUF	Instruct (4-bit)
DeepSeek-V3	V3-0324	link	—
	V3	link	—
DeepSeek-R1	R1-0528	link	—
	R1-0528-Qwen3-8B	link	link
	R1	link	—
	R1 Zero	link	—
	Distill Llama 3 8 B	link	link
	Distill Llama 3.3 70 B	link	link
	Distill Qwen 2.5 1.5 B	link	link
	Distill Qwen 2.5 7 B	link	link
	Distill Qwen 2.5 14 B	link	link
	Distill Qwen 2.5 32 B	link	link

Llama models:

Model	Variant	GGUF	Instruct (4-bit)
Llama 4	Scout 17 B-16 E	link	link
	Maverick 17 B-128 E	link	—
Llama 3.3	70 B	link	link
Llama 3.2	1 B	link	link
	3 B	link	link
	11 B Vision	—	link
	90 B Vision	—	link
Llama 3.1	8 B	link	link
	70 B	—	link
	405 B	—	link
Llama 3	8 B	—	link
	70 B	—	link
Llama 2	7 B	—	link
	13 B	—	link
CodeLlama	7 B	—	link
	13 B	—	link
	34 B	—	link

Gemma models:

Model	Variant	GGUF	Instruct (4-bit)
Gemma 3n	E2B	​link	link
	E4B	link	ink
Gemma 3	1 B	link	link
	4 B	link	link
	12 B	link	link
	27 B	link	link
MedGemma	4 B (vision)	link	link
	27 B (vision)	link	link
Gemma 2	2 B	link	link
	9 B	—	link
	27 B	—	link

Qwen models:

Model	Variant	GGUF	Instruct (4-bit)
Qwen 3	0.6 B	link	link
	1.7 B	link	link
	4 B	link	link
	8 B	link	link
	14 B	link	link
	30 B-A3B	link	link
	32 B	link	link
	235 B-A22B	link	—
Qwen 2.5 Omni	3 B	link	—
	7 B	link	—
Qwen 2.5 VL	3 B	link	link
	7 B	link	link
	32 B	link	link
	72 B	link	link
Qwen 2.5	0.5 B	—	link
	1.5 B	—	link
	3 B	—	link
	7 B	—	link
	14 B	—	link
	32 B	—	link
	72 B	—	link
Qwen 2.5 Coder (128 K)	0.5 B	link	link
	1.5 B	link	link
	3 B	link	link
	7 B	link	link
	14 B	link	link
	32 B	link	link
QwQ	32 B	link	link
QVQ (preview)	72 B	—	link
Qwen 2 (chat)	1.5 B	—	link
	7 B	—	link
	72 B	—	link
Qwen 2 VL	2 B	—	link
	7 B	—	link
	72 B	—	link

Mistral models:

Model	Variant	GGUF	Instruct (4-bit)
Mistral Small	3.2-24 B (2506)	link	link
	3.1-24 B (2503)	link	link
	3-24 B (2501)	link	link
Magistral	Small-24 B (2506)	link	link
Devstral	Small-24 B (2507)	link	link
	Small-24 B (2505)	link	link
Pixtral	12 B (2409)	—	link
Mistral Small	2409-22 B	—	link
Mistral NeMo	12 B (2407)	link	link
Mistral Large	2407	—	link
Mistral 7 B	v0.3	—	link
	v0.2	—	link
Mixtral	8 × 7 B	—	link

Phi models:

Model	Variant	GGUF	Instruct (4-bit)
Phi-4	Reasoning-plus	link	link
	Reasoning	link	link
	Mini-Reasoning	link	link
	Phi-4 (instruct)	link	link
	mini (instruct)	link	link
Phi-3.5	mini	—	link
Phi-3	mini	—	link
	medium	—	link

Other (Orpheus, Smol, Llava etc.) models:

Model	Variant	GGUF	Instruct (4-bit)
Hunyuan	A13B	link	—
Orpheus	0.1-ft (3B)	link	link
LLava	1.5 (7 B)	—	link
	1.6 Mistral (7 B)	—	link
TinyLlama	Chat	—	link
SmolLM 2	135 M	link	link
	360 M	link	link
	1.7 B	link	link
Zephyr-SFT	7 B	—	link
Yi	6 B (v1.5)	—	link
	6 B (v1.0)	—	link
	34 B (chat)	—	link
	34 B (base)	—	link

• Instruct 16-bit

16-bit and 8-bit Instruct models are used for inference or fine-tuning:

New models:

Model	Variant	Instruct (16-bit)
Gemma 3n	E2B	link
	E4B	link
DeepSeek-R1-0528	R1-0528-Qwen3-8B	link
	R1-0528	link
Mistral	Small 3.2 24B (2506)	link
	Small 3.1 24B (2503)	link
	Small 3.0 24B (2501)	link
	Magistral Small (2506)	link
Qwen 3	0.6 B	link
	1.7 B	link
	4 B	link
	8 B	link
	14 B	link
	30B-A3B	link
	32 B	link
	235B-A22B	link
Llama 4	Scout 17B-16E	link
	Maverick 17B-128E	link
Qwen 2.5 Omni	3 B	link
	7 B	link
Phi-4	Reasoning-plus	link
	Reasoning	link

DeepSeek models

Model	Variant	Instruct (16-bit)
DeepSeek-V3	V3-0324	link
	V3	link
DeepSeek-R1	R1-0528	link
	R1-0528-Qwen3-8B	link
	R1	link
	R1 Zero	link
	Distill Llama 3 8B	link
	Distill Llama 3.3 70B	link
	Distill Qwen 2.5 1.5B	link
	Distill Qwen 2.5 7B	link
	Distill Qwen 2.5 14B	link
	Distill Qwen 2.5 32B	link

Llama models

Family	Variant	Instruct (16-bit)
Llama 4	Scout 17B-16E	link
	Maverick 17B-128E	link
Llama 3.3	70 B	link
Llama 3.2	1 B	link
	3 B	link
	11 B Vision	link
	90 B Vision	link
Llama 3.1	8 B	link
	70 B	link
	405 B	link
Llama 3	8 B	link
	70 B	link
Llama 2	7 B	link

Gemma models:

Model	Variant	Instruct (16-bit)
Gemma 3n	E2B	link
	E4B	link
Gemma 3	1 B	link
	4 B	link
	12 B	link
	27 B	link
Gemma 2	2 B	link
	9 B	link
	27 B	link

Qwen models:

Family	Variant	Instruct (16-bit)
Qwen 3	0.6 B	link
	1.7 B	link
	4 B	link
	8 B	link
	14 B	link
	30B-A3B	link
	32 B	link
	235B-A22B	link
Qwen 2.5 Omni	3 B	link
	7 B	link
Qwen 2.5 VL	3 B	link
	7 B	link
	32 B	link
	72 B	link
Qwen 2.5	0.5 B	link
	1.5 B	link
	3 B	link
	7 B	link
	14 B	link
	32 B	link
	72 B	link
Qwen 2.5 Coder 128 K	0.5 B	link
	1.5 B	link
	3 B	link
	7 B	link
	14 B	link
	32 B	link
QwQ	32 B	link
QVQ (preview)	72 B	—
Qwen 2 (Chat)	1.5 B	link
	7 B	link
	72 B	link
Qwen 2 VL	2 B	link
	7 B	link
	72 B	link

Mistral models:

Model	Variant	Instruct (16-bit)
Mistral	Small 2409-22B	link
Mistral	Large 2407	link
Mistral	7B v0.3	link
Mistral	7B v0.2	link
Pixtral	12B 2409	link
Mixtral	8×7B	link
Mistral NeMo	12B 2407	link
Devstral	Small 2505	link

Phi models:

Model	Variant	Instruct (16-bit)
Phi-4	Reasoning-plus	link
	Reasoning	link
	Phi-4 (core)	link
	Mini-Reasoning	link
	Mini	link
Phi-3.5	Mini	link
Phi-3	Mini	link
	Medium	link

Text-to-Speech (TTS) models:

Model	Instruct (16-bit)
Orpheus-3B (v0.1 ft)	link
Orpheus-3B (v0.1 pt)	link
Sesame-CSM 1B	link
Whisper Large V3 (STT)	link
Llasa-TTS 1B	link
Spark-TTS 0.5B	link
Oute-TTS 1B	link

• Base 4 + 16-bit

Base models are usually used for fine-tuning purposes:

New models:

Model	Variant	Base (16-bit)	Base (4-bit)
Gemma 3n	E2B	link	link
	E4B	link	link
Qwen 3	0.6 B	link	link
	1.7 B	link	link
	4 B	link	link
	8 B	link	link
	14 B	link	link
	30B-A3B	link	link
Llama 4	Scout 17B 16E	link	link
	Maverick 17B 128E	link	—

Llama models:

Model	Variant	Base (16-bit)	Base (4-bit)
Llama 4	Scout 17B 16E	link	—
	Maverick 17B 128E	link	—
Llama 3.3	70 B	link	—
Llama 3.2	1 B	link	—
	3 B	link	—
	11 B Vision	link	—
	90 B Vision	link	—
Llama 3.1	8 B	link	—
	70 B	link	—
Llama 3	8 B	link	link
Llama 2	7 B	link	link
	13 B	link	link

Qwen models:

Model	Variant	Base (16-bit)	Base (4-bit)
Qwen 3	0.6 B	link	link
	1.7 B	link	link
	4 B	link	link
	8 B	link	link
	14 B	link	link
	30B-A3B	link	link
Qwen 2.5	0.5 B	link	link
	1.5 B	link	link
	3 B	link	link
	7 B	link	link
	14 B	link	link
	32 B	link	link
	72 B	link	link
Qwen 2	1.5 B	link	link
	7 B	link	link

Llama models:

Model	Variant	Base (16-bit)	Base (4-bit)
Llama 4	Scout 17B 16E	link	—
	Maverick 17B 128E	link	—
Llama 3.3	70 B	link	—
Llama 3.2	1 B	link	—
	3 B	link	—
	11 B Vision	link	—
	90 B Vision	link	—
Llama 3.1	8 B	link	—
	70 B	link	—
Llama 3	8 B	link	link
Llama 2	7 B	link	link
	13 B	link	link

Gemma models

Model	Variant	Base (16-bit)	Base (4-bit)
Gemma 3	1 B	link	link
	4 B	link	link
	12 B	link	link
	27 B	link	link
Gemma 2	2 B	link	—
	9 B	link	—
	27 B	link	—

Mistral models:

Model	Variant	Base (16-bit)	Base (4-bit)
Mistral	Small 24B 2501	link	—
	NeMo 12B 2407	link	—
	7B v0.3	link	link
	7B v0.2	link	link
	Pixtral 12B 2409	link	—

Other (TTS, TinyLlama) models:

Model	Variant	Base (16-bit)	Base (4-bit)
TinyLlama	1.1 B (Base)	link	link
Orpheus-3b	0.1-pretrained	link	link