Highest Quality Fine-Tunes

Top fine-tuned models ranked by our quality score — combining downloads, likes, documentation, and trending signals.

Last updated April 3, 2026 · Updated daily

Kokoro-82M by hexgrad holds the #1 position with 63 score, ahead of WanVideo_comfy at 60.

The top 10 is dominated by hexgrad, Kijai, unsloth. This is the first snapshot — future updates will track position changes and emerging trends.

The gap between #1 and #140 is 63 vs 40 score, showing a relatively competitive field.

🥇new

Kokoro-82M

Silver63

hexgrad · TTS · from yl4579/StyleTTS2-LJSpeech

63

score

🥈new

WanVideo_comfy

Silver60

Kijai · Other · from Wan-AI/Wan2.1-VACE-1.3B

60

score

🥉new

Qwen3.5-35B-A3B-GGUF

Silver53

unsloth · VLM · from Qwen/Qwen3.5-35B-A3B

53

score

4new

Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled

Silver53

Jackrong · VLM · from Qwen/Qwen3.5-27B

53

score

5new

roberta-base-squad2

Silver52

deepset · LLM - General · from FacebookAI/roberta-base

52

score

6new

Qwen3.5-35B-A3B-Uncensored-HauhauCS-Aggressive

Silver52

HauhauCS · VLM · from Qwen/Qwen3.5-35B-A3B

52

score

7new

personaplex-7b-v1

Silver52

nvidia · Audio · from kyutai/moshiko-pytorch-bf16

52

score

8new

all-MiniLM-L12-v2

Silver52

sentence-transformers · Embeddings · from microsoft/MiniLM-L12-H384-uncased

52

score

9new

VibeVoice-Realtime-0.5B

Silver51

microsoft · TTS · from Qwen/Qwen2.5-0.5B

51

score

10new

Qwen3.5-9B-GGUF

Silver51

unsloth · VLM · from Qwen/Qwen3.5-9B

51

score

11new

Wan2.2-Distill-Loras

Silver50

lightx2v · Image-to-Video · from adapter:Wan-AI/Wan2.2-I2V-A14B

50

score

12new

Qwen3.5-27B-GGUF

Silver50

unsloth · VLM · from Qwen/Qwen3.5-27B

50

score

13new

Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled-GGUF

Silver50

Jackrong · VLM · from Qwen/Qwen3.5-27B

50

score

14new

fairface_age_image_detection

Silver50

dima806 · Image Classification · from google/vit-base-patch16-224-in21k

50

score

15new

WanVideo_comfy_fp8_scaled

Bronze49

Kijai · Other · from Wan-AI/Wan2.1-VACE-1.3B

49

score

16new

dolphin-2.9.1-yi-1.5-34b

Bronze49

dphn · Code Generation (Uncensored) · from 01-ai/Yi-1.5-34B

49

score

17new

zephyr-7b-beta

Bronze49

HuggingFaceH4 · Chat & Assistants · from mistralai/Mistral-7B-v0.1

49

score

18new

Qwen-Image-Lightning

Bronze49

lightx2v · Image Generation · from adapter:Qwen/Qwen-Image

49

score

19new

Nanonets-OCR2-3B

Bronze48

nanonets · VLM · from Qwen/Qwen2.5-VL-3B-Instruct

48

score

20new

PaddleOCR-VL-1.5

Bronze48

PaddlePaddle · VLM · from baidu/ERNIE-4.5-0.3B-Paddle

48

score

21new

ms-marco-MiniLM-L12-v2

Bronze48

cross-encoder · Other · from microsoft/MiniLM-L12-H384-uncased

48

score

22new

gpt-oss-20b-GGUF

Bronze48

unsloth · LLM - General · from openai/gpt-oss-20b

48

score

23new

Qwen3-Coder-Next-GGUF

Bronze48

unsloth · Code Generation · from Qwen/Qwen3-Coder-Next

48

score

24new

xlm-roberta-base-language-detection

Bronze48

papluca · Other · from FacebookAI/xlm-roberta-base

48

score

25new

all-MiniLM-L6-v2

Bronze48

Xenova · Embeddings · from sentence-transformers/all-MiniLM-L6-v2

48

score

26new

pubmedbert-base-embeddings

Bronze48

NeuML · Embeddings · from microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract-fulltext

48

score

27new

multilingual-sentiment-analysis

Bronze48

tabularisai · Other · from distilbert/distilbert-base-multilingual-cased

48

score

28new

animagine-xl-4.0

Bronze47

cagliostrolab · Image Generation · from stabilityai/stable-diffusion-xl-base-1.0

47

score

29new

Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled-v2-GGUF

Bronze47

Jackrong · VLM · from adapter:Qwen/Qwen3.5-27B

47

score

30new

OpenHermes-2.5-Mistral-7B

Bronze47

teknium · Chat & Assistants · from mistralai/Mistral-7B-v0.1

47

score

31new

Qwen3.5-4B-GGUF

Bronze47

unsloth · VLM · from Qwen/Qwen3.5-4B

47

score

32new

gemma-2-2b-it-GGUF

Bronze47

bartowski · LLM - General · from google/gemma-2-2b-it

47

score

33new

GLM-4.7-Flash-GGUF

Bronze47

unsloth · LLM - General · from zai-org/GLM-4.7-Flash

47

score

34new

Llama-3.2-3B-Instruct-GGUF

Bronze47

bartowski · Chat & Assistants · from meta-llama/Llama-3.2-3B-Instruct

47

score

35new

Meta-Llama-3.1-8B-Instruct-GGUF

Bronze46

bartowski · Chat & Assistants · from meta-llama/Llama-3.1-8B-Instruct

46

score

36new

Hyper-SD

Bronze46

ByteDance · Image Generation · from adapter:black-forest-labs/FLUX.1-dev

46

score

37new

FLUX.1-dev-gguf

Bronze46

city96 · Image Generation · from black-forest-labs/FLUX.1-dev

46

score

38new

Molmo2-8B

Bronze46

allenai · VLM · from Qwen/Qwen3-8B

46

score

39new

RolmOCR

Bronze46

reducto · VLM · from Qwen/Qwen2.5-VL-7B-Instruct

46

score

40new

Nanonets-OCR-s

Bronze46

nanonets · VLM · from Qwen/Qwen2.5-VL-3B-Instruct

46

score

41new

gte-reranker-modernbert-base

Bronze46

Alibaba-NLP · Other · from answerdotai/ModernBERT-base

46

score

42new

mdeberta-v3-base-squad2

Bronze46

timpal0l · LLM - General · from microsoft/mdeberta-v3-base

46

score

43new

SmolDocling-256M-preview

Bronze46

docling-project · VLM · from HuggingFaceTB/SmolVLM-256M-Instruct

46

score

44new

Wan2.1-VAE-upscale2x

Bronze46

spacepxl · Other · from Wan-AI/Wan2.1-T2V-14B

46

score

45new

animagine-xl-3.0

Bronze46

cagliostrolab · Image Generation · from Linaqruf/animagine-xl-2.0

46

score

46new

InternVL2-2B

Bronze46

OpenGVLab · VLM · from merge:internlm/internlm2-chat-1_8b

46

score

47new

Llama-2-7B-Chat-GGUF

Bronze46

TheBloke · Chat & Assistants · from meta-llama/Llama-2-7b-chat-hf

46

score

48new

Qwen-Edit-2509-Multiple-angles

Bronze46

dx8152 · Image-to-Image · from adapter:Qwen/Qwen-Image-Edit-2509

46

score

49new

InternVL2-8B

Bronze46

OpenGVLab · VLM · from merge:internlm/internlm2_5-7b-chat

46

score

50new

Qwen3-Coder-30B-A3B-Instruct-GGUF

Bronze46

unsloth · Code Generation · from Qwen/Qwen3-Coder-30B-A3B-Instruct

46

score

51new

facial_emotions_image_detection

Bronze46

dima806 · Image Classification · from google/vit-base-patch16-224-in21k

46

score

52new

Qwen3.5-122B-A10B-GGUF

Bronze46

unsloth · VLM · from Qwen/Qwen3.5-122B-A10B

46

score

53new

FHDR_Uncensored

Bronze46

kpsss34 · Image Generation (Uncensored) · from black-forest-labs/FLUX.1-dev

46

score

54new

Qwen-Image-Edit-2511-Lightning

Bronze46

lightx2v · Image-to-Image · from adapter:Qwen/Qwen-Image-Edit-2511

46

score

55new

Nemotron-3-Nano-30B-A3B-GGUF

Bronze45

unsloth · Code Generation · from nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16

45

score

55new

gpt-oss-120b-GGUF

Bronze43

ggml-org · Other · from openai/gpt-oss-120b

43

score

56new

Kokoro-82M-bf16

Bronze43

mlx-community · TTS · from yl4579/StyleTTS2-LJSpeech

43

score

56new

LTX-2.3-GGUF

Bronze45

unsloth · Image-to-Video · from Lightricks/LTX-2.3

45

score

57new

canary-qwen-2.5b

Bronze45

nvidia · Speech-to-Text · from Qwen/Qwen3-1.7B

45

score

58new

Qwen3.5-9B-Claude-4.6-Opus-Reasoning-Distilled-GGUF

Bronze45

Jackrong · Math & Reasoning · from adapter:Qwen/Qwen3.5-9B

45

score

59new

TableGPT2-7B

Bronze45

tablegpt · Other · from Qwen/Qwen2.5-7B

45

score

60new

llama-joycaption-alpha-two-hf-llava

Bronze45

fancyfeast · Other · from google/siglip-so400m-patch14-384

45

score

61new

stable-diffusion-xl-1.0-inpainting-0.1

Bronze45

diffusers · Image Generation · from stabilityai/stable-diffusion-xl-base-1.0

45

score

61new

turn-detector

Bronze43

livekit · Other · from Qwen/Qwen2.5-0.5B-Instruct

43

score

62new

olmOCR-2-7B-1025

Bronze45

allenai · VLM · from Qwen/Qwen2.5-VL-7B-Instruct

45

score

63new

deberta-v3-large-zeroshot-v2.0

Bronze45

MoritzLaurer · LLM - General · from microsoft/deberta-v3-large

45

score

64new

Kimi-K2.5-NVFP4

Bronze45

nvidia · Code Generation · from moonshotai/Kimi-K2.5

45

score

64new

Meta-Llama-3.1-8B-Instruct-bnb-4bit

Bronze43

unsloth · Chat & Assistants · from meta-llama/Llama-3.1-8B-Instruct

43

score

65new

BioClinical-ModernBERT-base

Bronze43

thomas-sounack · Code Generation · from answerdotai/ModernBERT-base

43

score

65new

Qwen-Image-Edit-2511-Multiple-Angles-LoRA

Bronze45

fal · Image-to-Image · from adapter:Qwen/Qwen-Image-Edit-2511

45

score

66new

Qwen3.5-9B-GGUF

Bronze43

lmstudio-community · Other · from Qwen/Qwen3.5-9B

43

score

66new

olmOCR-7B-0225-preview

Bronze44

allenai · VLM · from Qwen/Qwen2-VL-7B-Instruct

44

score

67new

Qwen3-Coder-Next-NVFP4

Bronze42

GadflyII · Code Generation · from Qwen/Qwen3-Coder-Next

42

score

67new

Qwen3.5-9B-Claude-4.6-Opus-Reasoning-Distilled-v2-GGUF

Bronze44

Jackrong · VLM · from adapter:Qwen/Qwen3.5-9B

44

score

68new

gpt-oss-20b-MXFP4-Q8

Bronze42

mlx-community · LLM - General · from openai/gpt-oss-20b

42

score

68new

InternVL2-1B

Bronze44

OpenGVLab · VLM · from merge:Qwen/Qwen2-0.5B-Instruct

44

score

69new

llama-3.3-70b-instruct-awq

Bronze42

casperhansen · Chat & Assistants · from meta-llama/Llama-3.1-70B

42

score

69new

Mistral-7B-Instruct-v0.2-GGUF

Bronze44

TheBloke · Chat & Assistants · from mistralai/Mistral-7B-Instruct-v0.2

44

score

70new

Illustrious-xl-early-release-v0

Bronze44

OnomaAIResearch · Image Generation · from KBlueLeaf/kohaku-xl-beta5

44

score

70new

parakeet-tdt-0.6b-v2

Bronze42

mlx-community · Speech-to-Text · from nvidia/parakeet-tdt-0.6b-v2

42

score

71new

lcm-lora-sdv1-5

Bronze44

latent-consistency · Image Generation · from adapter:runwayml/stable-diffusion-v1-5

44

score

71new

Wan2.1

Bronze42

DeepBeepMeep · Other · from Wan-AI/Wan2.1-T2V-1.3B

42

score

72new

jina-embeddings-v5-text-nano

Bronze42

jinaai · Embeddings · from EuroBERT/EuroBERT-210m

42

score

72new

Qwen3.5-397B-A17B-GGUF

Bronze44

unsloth · VLM · from Qwen/Qwen3.5-397B-A17B

44

score

73new

detr-doc-table-detection

Bronze42

TahaDouaji · Object Detection · from facebook/detr-resnet-50

42

score

73new

SeedVR2_comfyUI

Bronze44

numz · Other · from ByteDance-Seed/SeedVR2-3B

44

score

74new

Crow-9B-HERETIC-4.6

Bronze44

Crownelius · Other · from trohrbaugh/Qwen3.5-9B-heretic-v2

44

score

74new

Qwen3.5-35B-A3B-AWQ-4bit

Bronze42

cyankiwi · VLM · from Qwen/Qwen3.5-35B-A3B

42

score

75new

parakeet-tdt-0.6b-v3

Bronze42

mlx-community · Speech-to-Text · from nvidia/parakeet-tdt-0.6b-v3

42

score

75new

Hermes-3-Llama-3.1-8B

Bronze44

NousResearch · Role-Play & Characters · from meta-llama/Llama-3.1-8B

44

score

76new

nsfw_image_detector

Bronze44

Freepik · Image Classification · from timm/eva02_base_patch14_448.mim_in22k_ft_in22k_in1k

44

score

76new

ms-marco-TinyBERT-L2-v2

Bronze42

cross-encoder · Other · from nreimers/BERT-Tiny_L-2_H-128_A-2

42

score

77new

Qwen2.5-72B-Instruct-abliterated

Bronze42

huihui-ai · Chat & Assistants (Uncensored) · from Qwen/Qwen2.5-72B-Instruct

42

score

77new

Jan-v1-4B

Bronze44

janhq · Math & Reasoning · from Qwen/Qwen3-4B-Thinking-2507

44

score

78new

mmarco-mMiniLMv2-L12-H384-v1

Bronze44

cross-encoder · Other · from nreimers/mMiniLMv2-L12-H384-distilled-from-XLMR-Large

44

score

78new

Mistral-7B-Instruct-v0.2-AWQ

Bronze42

TheBloke · Chat & Assistants · from mistralai/Mistral-7B-Instruct-v0.2

42

score

79new

Qwen3.5-122B-A10B-NVFP4

Bronze42

Sehyo · VLM · from Qwen/Qwen3.5-122B-A10B

42

score

79new

vit-base-nsfw-detector

Bronze44

AdamCodd · Image Classification · from google/vit-base-patch16-384

44

score

80new

Vietnamese_Embedding

Bronze42

AITeamVN · Embeddings · from BAAI/bge-m3

42

score

80new

deberta-v3-base-prompt-injection-v2

Bronze44

protectai · Other · from microsoft/deberta-v3-base

44

score

81new

wav2vec2-large-robust-24-ft-age-gender

Bronze41

audeering · Audio · from facebook/wav2vec2-large-robust

41

score

81new

olmOCR-2-7B-1025-FP8

Bronze44

allenai · VLM · from Qwen/Qwen2.5-VL-7B-Instruct

44

score

82new

Phi-3.5-mini-instruct-GGUF

Bronze41

MaziyarPanahi · Role-Play & Characters · from microsoft/Phi-3.5-mini-instruct

41

score

82new

modernbert-embed-base

Bronze44

nomic-ai · Embeddings · from answerdotai/ModernBERT-base

44

score

83new

Qwen3.5-27B-AWQ

Bronze41

QuantTrio · VLM · from Qwen/Qwen3.5-27B

41

score

83new

Wan2.2-I2V-A14B-GGUF

Bronze44

QuantStack · Image-to-Video · from Wan-AI/Wan2.2-I2V-A14B

44

score

84new

bge-base-en-v1.5

Bronze41

Xenova · Embeddings · from BAAI/bge-base-en-v1.5

41

score

84new

Qwen3.5-0.8B-GGUF

Bronze44

unsloth · VLM · from Qwen/Qwen3.5-0.8B

44

score

85new

GLM-4.7-Flash-AWQ-4bit

Bronze41

cyankiwi · LLM - General · from zai-org/GLM-4.7-Flash

41

score

85new

Qwen-Image-Edit-2511-GGUF

Bronze44

unsloth · Image-to-Image · from Qwen/Qwen-Image-Edit-2511

44

score

86new

DeepSeek-R1-0528-NVFP4-v2

Bronze41

nvidia · Code Generation · from deepseek-ai/DeepSeek-R1-0528

41

score

86new

SmolVLM-500M-Instruct-GGUF

Bronze44

ggml-org · Other · from HuggingFaceTB/SmolVLM-500M-Instruct

44

score

87new

Qwen_Qwen3.5-35B-A3B-GGUF

Bronze41

bartowski · VLM · from Qwen/Qwen3.5-35B-A3B

41

score

87new

Llama-3.2-1B-Instruct-Q8_0-GGUF

Bronze44

hugging-quants · Chat & Assistants · from meta-llama/Llama-3.2-1B-Instruct

44

score

88new

DeepSeek-R1-GGUF

Bronze44

unsloth · Code Generation · from deepseek-ai/DeepSeek-R1

44

score

88new

Meta-Llama-3.1-8B-Instruct-FP8

Bronze41

RedHatAI · Chat & Assistants · from meta-llama/Llama-3.1-8B-Instruct

41

score

89new

jina-reranker-m0

Bronze43

jinaai · Other · from Qwen/Qwen2-VL-2B-Instruct

43

score

89new

cryptoNER

Bronze41

covalenthq · LLM - General · from FacebookAI/xlm-roberta-base

41

score

90new

watt-tool-8B

Bronze43

watt-ai · Other · from meta-llama/Llama-3.1-8B-Instruct

43

score

90new

Qwen2.5-Coder-7B-Instruct-GGUF

Bronze41

bartowski · Code Generation · from Qwen/Qwen2.5-Coder-7B-Instruct

41

score

91new

Qwen3-14B-Instruct

Bronze40

OpenPipe · Chat & Assistants · from Qwen/Qwen3-14B-Base

40

score

91new

indonesian-roberta-base-posp-tagger

Bronze43

w11wo · LLM - General · from flax-community/indonesian-roberta-base

43

score

92new

wai-nsfw-illustrious-v80-sdxl

Bronze40

John6666 · Image Generation (Uncensored) · from OnomaAIResearch/Illustrious-xl-early-release-v0

40

score

92new

Wav2Vec2-large-xlsr-hindi

Bronze43

theainerd · Speech-to-Text · from facebook/wav2vec2-large-xlsr-53

43

score

93new

DeepSeek-R1-0528-Qwen3-8B-GGUF

Bronze43

unsloth · LLM - General · from deepseek-ai/DeepSeek-R1-0528-Qwen3-8B

43

score

93new

DeepSeek-OCR-2

Bronze40

unsloth · VLM · from deepseek-ai/DeepSeek-OCR-2

40

score

94new

Baichuan-M2-32B

Bronze43

baichuan-inc · Chat & Assistants · from Qwen/Qwen2.5-32B

43

score

94new

Mistral-Small-24B-Instruct-2501-AWQ

Bronze40

stelterlab · Chat & Assistants · from mistralai/Mistral-Small-24B-Instruct-2501

40

score

95new

jina-reranker-v3

Bronze43

jinaai · Other · from Qwen/Qwen3-0.6B

43

score

95new

MedEmbed-base-v0.1

Bronze40

abhinand · Other · from BAAI/bge-base-en-v1.5

40

score

96new

Qwen3.5-35B-A3B-GGUF

Bronze40

lmstudio-community · Other · from Qwen/Qwen3.5-35B-A3B

40

score

96new

Wan2.2-I2V-A14B-GGUF

Bronze43

bullerwins · Image-to-Video · from Wan-AI/Wan2.2-I2V-A14B

43

score

97new

gemma-3-4b-it-GGUF

Bronze40

lmstudio-community · VLM · from google/gemma-3-4b-it

40

score

97new

Meta-Llama-3.1-8B-Instruct

Bronze43

unsloth · Chat & Assistants · from meta-llama/Llama-3.1-8B-Instruct

43

score

98new

distil-whisper-large-v3-ptbr

Bronze40

freds0 · Speech-to-Text · from distil-whisper/distil-large-v3

40

score

98new

Kimi-K2.5

Bronze43

mlx-community · Code Generation · from moonshotai/Kimi-K2.5

43

score

99new

Qwen3.5-2B-GGUF

Bronze43

unsloth · VLM · from Qwen/Qwen3.5-2B

43

score

99new

wav2vec2-large-xls-r-300m-Urdu

Bronze40

kingabzpro · Speech-to-Text · from facebook/wav2vec2-xls-r-300m

40

score

100new

Qwen3-VL-30B-A3B-Instruct-AWQ

Bronze43

QuantTrio · Chat & Assistants · from Qwen/Qwen3-VL-30B-A3B-Instruct

43

score

100new

mistral-7b-v0.3-bnb-4bit

Bronze40

unsloth · LLM - General · from mistralai/Mistral-7B-v0.3

40

score

About the Highest Quality Fine-Tunes Leaderboard

Top fine-tuned models ranked by our quality score — combining downloads, likes, documentation, and trending signals. This leaderboard tracks the top 140 fine-tuned models ranked by score, with daily snapshots to monitor how the rankings evolve over time.

Unlike HuggingFace's default model hub, Fine-Tune Catalog filters out pure quantizations and format conversions (repos that simply re-package a model as GGUF, AWQ, GPTQ, or EXL2 without any new training). Fine-tuned models published in quantized formats are still included — what matters is whether new training happened, not the output format.

Methodology

Rankings use our composite quality score (0-100) which combines download volume (40%), community likes (30%), documentation quality (15%), and trending momentum (15%). This balances popularity with substance.

Rankings are snapshotted daily at 6:00 AM UTC. Position changes shown on the leaderboard compare the current snapshot to the previous day's snapshot. All data is sourced directly from the HuggingFace Hub API and processed through our classification pipeline, which uses tag analysis, model card parsing, and naming pattern detection to identify genuine fine-tunes.

Data Sources

  • HuggingFace Hub API — download counts, likes, trending scores, model metadata, and README/model cards
  • Model card parsing — training datasets, training method (LoRA, DPO, SFT, etc.), framework, hardware, and hyperparameters extracted from README files
  • Tag classification — fine-tune detection via `base_model:finetune:*` and `base_model:quantized:*` HuggingFace tags, plus naming pattern analysis

Who Is This For?

This leaderboard is designed for AI engineers choosing pre-trained models, researchers benchmarking fine-tuning approaches, and developers building AI-powered applications who need to identify the most battle-tested and community-validated models.

Whether you're a beginner exploring what's possible with fine-tuned AI models or an experienced ML engineer looking for the best starting point for your next project, these rankings give you a data-driven way to find the highest quality models without having to wade through thousands of quantizations, format conversions, and abandoned repositories on HuggingFace.

Update Schedule

This leaderboard was last updated on April 3, 2026. Rankings are refreshed daily with the latest download counts, likes, and trending data from HuggingFace. Historical snapshots are preserved to track trends over time — you can see which models are gaining traction and which are losing momentum.

Explore more rankings: Most Downloaded Fine-Tunes · Most Liked Fine-Tunes · Top Uncensored Models · Top LLM Fine-Tunes · Top VLM Fine-Tunes · Top Image Generation Fine-Tunes and 16 more leaderboards