GitHub‑OSS‑Trends

Heißeste Open-Source-Projekte der letzten 24 Stunden

public-apis/public-apis · ⭐ 463,455 freeCodeCamp/freeCodeCamp · ⭐ 454,225 EbookFoundation/free-programming-books · ⭐ 394,659 openclaw/openclaw · ⭐ 386,598 nilbuild/developer-roadmap · ⭐ 364,786 vinta/awesome-python · ⭐ 314,559 awesome-selfhosted/awesome-selfhosted · ⭐ 313,366 practical-tutorials/project-based-learning · ⭐ 279,676 public-apis/public-apis · ⭐ 463,455 freeCodeCamp/freeCodeCamp · ⭐ 454,225 EbookFoundation/free-programming-books · ⭐ 394,659 openclaw/openclaw · ⭐ 386,598 nilbuild/developer-roadmap · ⭐ 364,786 vinta/awesome-python · ⭐ 314,559 awesome-selfhosted/awesome-selfhosted · ⭐ 313,366 practical-tutorials/project-based-learning · ⭐ 279,676

Live source: HF Trending Snapshot source: GitHub grouped snapshot Estimated: Heat / VRAM

Angesagte Modelle

Δ 24 h

Die Standardreihenfolge entspricht der Hugging-Face-Trendliste (7-Tage-Likes).

Nur kommerziell nutzbare Lizenzen (z. B. Apache 2.0 / MIT) New listings only

Qwen3.8-27B New

Qwen

Δ 24 h +67%

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-text

HF

Model Card

Hippo Heat 96

See model card
Qwen3.8-27B-GGUF New

unsloth

Δ 24 h +HF Hot

Lizenz — Other LLM

VRAM

—

HF-Task

—

HF

Model Card

Hippo Heat 85

See model card
Qwen3.8-2.4T-A95B New

Qwen

Δ 24 h +48%

Lizenz — Other LLM

VRAM

—

HF-Task

text-generation

HF

Model Card

Hippo Heat 74

See model card
LTX-2.5 New Hot

Lightricks

Δ 24 h +68%

Lizenz — Other Vision

VRAM

—

HF-Task

image-to-video

HF

Model Card

Hippo Heat 97

See model card
MiniMax-Music3 New

MiniMaxAI

Δ 24 h +48%

Lizenz — Other Audio

VRAM

—

HF-Task

text-to-audio

HF

Model Card

Hippo Heat 75

See model card
Muse-Glimmer-30B New

meta-models

Δ 24 h +66%

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-text

HF

Model Card

Hippo Heat 94

See model card
MiniMax-H3 New

MiniMaxAI

Δ 24 h +HF Hot

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-video

HF

Model Card

Hippo Heat 21

See model card
DeepSeek-V4-Pro-0813 New

deepseek-ai

Δ 24 h +53%

Lizenz — Other LLM

VRAM

—

HF-Task

text-generation

HF

Model Card

Hippo Heat 80

See model card
Qwen3.8-27B-FP8 New Hot

Qwen

Δ 24 h +68%

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-text

HF

Model Card

Hippo Heat 97

See model card
Qwen3.8-27B-Uncensored-FP8 New

orcarouter

Δ 24 h +50%

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-text

HF

Model Card

Hippo Heat 77

See model card
DeepSeek-V4-Flash-0731 New

deepseek-ai

Δ 24 h +HF Hot

Lizenz — Other LLM

VRAM

—

HF-Task

text-generation

HF

Model Card

Hippo Heat 21

See model card
Qwen3.8-27B-Uncensored-GGUF New

JonathanColetti

Δ 24 h +67%

Lizenz — Other LLM

VRAM

—

HF-Task

text-generation

HF

Model Card

Hippo Heat 96

See model card
Minimax-h3-Turbo New

lightx2v

Δ 24 h +65%

Lizenz — Other Vision

VRAM

—

HF-Task

image-to-video

HF

Model Card

Hippo Heat 93

See model card
NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4 New

nvidia

Δ 24 h +64%

Lizenz — Other LLM

VRAM

—

HF-Task

text-generation

HF

Model Card

Hippo Heat 92

See model card
Kimi-K3 New

moonshotai

Δ 24 h +HF Hot

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-text

HF

Model Card

Hippo Heat 21

See model card
Qwen3.8-27B-NVFP4 New

unsloth

Δ 24 h +67%

Lizenz — Other LLM

VRAM

—

HF-Task

—

HF

Model Card

Hippo Heat 96

See model card
Muse-Glimmer-30B-GGUF New Hot

unsloth

Δ 24 h +71%

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-text

HF

Model Card

Hippo Heat 96

See model card
Anima-2.9B New

Gazingstars123

Δ 24 h +52%

Lizenz — Other Vision

VRAM

—

HF-Task

text-to-image

HF

Model Card

Hippo Heat 79

See model card
Ling-3.0-tiny New

inclusionAI

Δ 24 h +46%

Lizenz — Other LLM

VRAM

—

HF-Task

text-generation

HF

Model Card

Hippo Heat 71

See model card
Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-NEO-MAX-MTP-GGUF New

DavidAU

Δ 24 h +HF Hot

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-text

HF

Model Card

Hippo Heat 21

See model card
Qwen3.8-2.4T-A95B-FP8 New

Qwen

Δ 24 h +49%

Lizenz — Other LLM

VRAM

—

HF-Task

text-generation

HF

Model Card

Hippo Heat 76

See model card
MiniMax-H3 New

Comfy-Org

Δ 24 h +HF Hot

Lizenz — Other LLM

VRAM

—

HF-Task

—

HF

Model Card

Hippo Heat 21

See model card
dots3-note-prev New

dots-studio

Δ 24 h +34%

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-text

HF

Model Card

Hippo Heat 53

See model card
Qwen-Fixed-Chat-Templates New

froggeric

Δ 24 h —

Lizenz — Other LLM

VRAM

—

HF-Task

—

HF

Model Card

Hippo Heat 0

See model card
MiniMax-Music-3 New

Comfy-Org

Δ 24 h +65%

Lizenz — Other LLM

VRAM

—

HF-Task

—

HF

Model Card

Hippo Heat 93

See model card
NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16 New

nvidia

Δ 24 h +58%

Lizenz — Other LLM

VRAM

—

HF-Task

text-generation

HF

Model Card

Hippo Heat 86

See model card
LFM2.5-VL-3B New

LiquidAI

Δ 24 h +46%

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-text

HF

Model Card

Hippo Heat 71

See model card
MiniMax-H3-Realism-People-LoRA New

fal

Δ 24 h +51%

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-video

HF

Model Card

Hippo Heat 78

See model card
10Eros-Max New

TenStrip

Δ 24 h —

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-video

HF

Model Card

Hippo Heat 0

See model card
MiniMax-H3-Turbo-Lora New

larryvrh

Δ 24 h —

Lizenz — Other Vision

VRAM

—

HF-Task

text-to-video

HF

Model Card

Hippo Heat 0

See model card
LFM2.5-2.6B New

LiquidAI

Δ 24 h +62%

Lizenz — Other LLM

VRAM

—

HF-Task

text-generation

HF

Model Card

Hippo Heat 90

See model card
Qwen3.8-27B-ABLITERATED-GGUF New

Blackfrost-AI

Δ 24 h +60%

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-text

HF

Model Card

Hippo Heat 88

See model card
Qwen3.8-27B-Ridge-GGUF New

empero-ai

Δ 24 h +48%

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-text

HF

Model Card

Hippo Heat 74

See model card
needle2 New

Cactus-Compute

Δ 24 h +48%

Lizenz — Other LLM

VRAM

—

HF-Task

text-generation

HF

Model Card

Hippo Heat 75

See model card
Qwen3.6-27B-Uncensored-HauhauCS-Aggressive

HauhauCS

Δ 24 h +66%

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-text

HF

Model Card

Hippo Heat 94

See model card
Qwen3.8-27B-AEON-ULTIMATE-UNCENSORED-BF16 New

AEON-7

Δ 24 h +40%

Lizenz — Other LLM

VRAM

—

HF-Task

text-generation

HF

Model Card

Hippo Heat 62

See model card
IndexTTS-2.5 New

IndexTeam

Δ 24 h +43%

Lizenz — Other Audio

VRAM

—

HF-Task

text-to-speech

HF

Model Card

Hippo Heat 67

See model card
Muse-Glimmer-30B-GGUF New Hot

meta-models

Δ 24 h +67%

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-text

HF

Model Card

Hippo Heat 96

See model card
Qwen3.8-27B-Heretic-Abliterated-Uncensored-GGUF New

0bserverx

Δ 24 h +60%

Lizenz — Other LLM

VRAM

—

HF-Task

text-generation

HF

Model Card

Hippo Heat 88

See model card
North-Micro-Vision-Instruct New

CohereLabs

Δ 24 h +51%

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-text

HF

Model Card

Hippo Heat 78

See model card
Qwen3.8-9B New

empero-ai

Δ 24 h +33%

Lizenz — Other LLM

VRAM

—

HF-Task

text-generation

HF

Model Card

Hippo Heat 51

See model card
Qwen3.8-2.4T-A95B-GGUF New

unsloth

Δ 24 h +50%

Lizenz — Other LLM

VRAM

—

HF-Task

text-generation

HF

Model Card

Hippo Heat 77

See model card
Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-MTP-GGUF

HauhauCS

Δ 24 h —

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-text

HF

Model Card

Hippo Heat 0

See model card
Qwen3.8-27B-GGUF New

AtomicChat

Δ 24 h +57%

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-text

HF

Model Card

Hippo Heat 85

See model card
Qwen3.8-27B-Cold-Fusion-GAIN-V1.1 New

DavidAU

Δ 24 h +17%

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-text

HF

Model Card

Hippo Heat 27

See model card
MiniMax-H3_comfy New

Kijai

Δ 24 h —

Lizenz — Other LLM

VRAM

—

HF-Task

—

HF

Model Card

Hippo Heat 0

See model card
MiniMax-H3-GGUF New

unsloth

Δ 24 h +64%

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-video

HF

Model Card

Hippo Heat 92

See model card
Qwen3.8-27B-Abliterated-MLX New

PocketAiHub

Δ 24 h —

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-text

HF

Model Card

Hippo Heat 0

See model card
Qwen3.8-27B-Uncensored-MLX New

orcarouter

Δ 24 h —

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-text

HF

Model Card

Hippo Heat 0

See model card
Huihui-Qwen3.8-27B-abliterated New

huihui-ai

Δ 24 h +38%

Lizenz — Other Vision

VRAM

—

HF-Task

image-text-to-text

HF

Model Card

Hippo Heat 59

See model card

Gruppierter Repository-Vergleich

Jeder Abschnitt listet Repos derselben Klasse (z. B. Agent: OpenClaw, NanoBot, PicoClaw). Die Spaltenbedeutungen sind in allen Gruppen gleich. „Commits (30 T.)“ folgt dem rollierenden Fenster des Scrapers.

Snapshot: 18.08.2026

Datenquellen und Methodik

Public formula (within each group): Score = 100 × [0.30·Stars + 0.15·Forks + 0.30·Commits30d + 0.15·Contributors + 0.05·(1-Issues) + 0.05·(1-PRs)], after per-group min-max normalization.

Agent

Rank	Repository	Score	Stars	Forks	Commits (30 T.)	Mitwirkende	Open issues	Open PRs
1	openclaw/openclaw	96,6	386.598	81.238	9.781	3.108	5.713	2
2	NousResearch/hermes-agent	58,8	232.196	46.269	7.299	2.896	32.879	5
3	deepseek-ai/deepseek-harness	42,7	154.829	16.019	9.158	33	0	0
4	langchain-ai/langchain	29,8	144.444	24.045	176	3.714	410	84
5	affaan-m/ECC	29,6	240.792	36.523	156	318	132	90
6	Significant-Gravitas/AutoGPT	26,4	186.662	46.061	105	839	506	218
7	Shubhamsaboo/awesome-llm-apps	13,1	133.031	19.574	75	106	13	7
8	firecrawl/firecrawl	12,6	168.780	9.437	138	170	521	454
9	langflow-ai/langflow	10,3	153.393	9.879	364	371	967	753
10	farion1231/cc-switch	5,7	127.961	8.756	192	211	2.235	758

Multi-Agent

Rank	Repository	Score	Stars	Forks	Commits (30 T.)	Mitwirkende	Open issues	Open PRs
1	lobehub/lobehub	78,1	81.792	15.806	712	352	755	458
2	TauricResearch/TradingAgents	50,6	98.733	19.030	0	20	365	183
3	danny-avila/LibreChat	40,6	42.167	8.727	313	391	703	347
4	openai/openai-agents-python	40,4	28.732	4.527	377	354	12	3
5	zhayujie/CowAgent	37,4	46.542	10.319	199	144	33	14
6	FoundationAgents/MetaGPT	36	69.874	8.885	0	148	129	113
7	Yeachan-Heo/oh-my-claudecode	24,4	38.625	3.471	124	140	2	0
8	wshobson/agents	18	38.881	4.141	16	79	16	7
9	OpenBMB/ChatDev	12,6	34.014	4.245	3	19	67	21
10	agentscope-ai/agentscope	11,8	29.006	3.369	51	97	367	111

Inferenz

Rank	Repository	Score	Stars	Forks	Commits (30 T.)	Mitwirkende	Open issues	Open PRs
1	ggml-org/llama.cpp	73,2	124.446	21.842	414	1.912	2.103	1
2	BerriAI/litellm	66,6	56.603	10.672	2.347	1.595	4.934	3
3	earendil-works/pi	50,3	92.672	11.476	731	281	140	48
4	sgl-project/sglang	47	31.995	7.971	1.358	1.774	4.927	4
5	microsoft/onnxruntime	19,5	21.399	4.131	265	966	1.580	764
6	k2-fsa/sherpa-onnx	13,2	14.228	1.632	70	230	615	53
7	GeeeekExplorer/nano-vllm	12	15.040	2.470	0	15	81	51
8	abetlen/llama-cpp-python	10,8	10.560	1.441	7	195	680	82
9	h2oai/h2ogpt	10,7	11.972	1.302	0	78	328	36
10	niedev/RTranslator	10	10.330	938	0	12	26	1

LLM-Bibliothek

Rank	Repository	Score	Stars	Forks	Commits (30 T.)	Mitwirkende	Open issues	Open PRs
1	huggingface/transformers	71,5	164.211	34.264	263	4.067	2.395	1
2	unslothai/unsloth	44,9	73.350	6.614	1.215	276	1.278	393
3	fighting41love/funNLP	26,8	82.521	15.265	0	13	53	9
4	router-for-me/CLIProxyAPI	21,2	47.679	7.372	387	239	388	337
5	asgeirtj/system_prompts_leaks	20,9	63.097	10.357	50	25	56	35
6	sickn33/antigravity-awesome-skills	20,5	45.082	6.597	211	327	0	0
7	labmlai/annotated_deep_learning_paper_implementations	19,3	67.312	6.743	0	42	34	6
8	ComposioHQ/awesome-claude-skills	18,8	72.700	8.310	2	29	1.283	1
9	langfuse/langfuse	18,3	33.298	3.583	578	197	784	483
10	huggingface/pytorch-image-models	12,5	37.071	5.186	18	198	61	19

Trainings-Framework

Rank	Repository	Score	Stars	Forks	Commits (30 T.)	Mitwirkende	Open issues	Open PRs
1	pytorch/pytorch	70,7	102.450	28.900	1.401	6.856	17.078	3
2	vllm-project/vllm	54,4	89.321	20.830	1.200	3.223	6.678	4
3	AUTOMATIC1111/stable-diffusion-webui	48,8	164.557	30.553	0	638	2.502	80
4	tensorflow/models	30,7	77.652	44.897	16	959	1.274	198
5	Comfy-Org/ComfyUI	30,4	128.195	15.095	179	338	4.619	518
6	rasbt/LLMs-from-scratch	25,5	102.869	15.763	4	68	2	0
7	keras-team/keras	20,2	64.239	19.750	75	1.551	224	86
8	ultralytics/ultralytics	18,8	60.714	11.587	345	424	167	119
9	ultralytics/yolov5	14,8	57.894	17.473	31	349	31	1
10	CorentinJ/Real-Time-Voice-Cloning	10,5	60.091	9.398	0	20	177	8

Serving

Rank	Repository	Score	Stars	Forks	Commits (30 T.)	Mitwirkende	Open issues	Open PRs
1	open-webui/open-webui	93,3	149.090	21.716	359	863	356	165
2	Mintplex-Labs/anything-llm	30,8	64.859	7.144	60	224	332	31
3	AlexsJones/llmfit	24,5	32.490	2.007	128	94	67	21
4	1Panel-dev/1Panel	23	36.568	3.299	90	128	289	14
5	chatchat-space/Langchain-Chatchat	20	38.546	6.262	0	173	26	20
6	janhq/jan	18,6	44.038	2.972	46	181	452	89
7	PDFMathTranslate/PDFMathTranslate	14,5	36.207	3.231	0	57	147	23
8	lm-sys/FastChat	11,7	39.513	4.785	0	246	1.039	144
9	lss233/kirara-ai	10,6	18.945	1.835	0	48	4	1
10	Zackriya-Solutions/meetily	7,7	29.339	3.124	0	12	348	134

Evaluation

Rank	Repository	Score	Stars	Forks	Commits (30 T.)	Mitwirkende	Open issues	Open PRs
1	confident-ai/deepeval	64,6	17.655	1.820	202	329	467	221
2	bitnami/charts	55,9	10.399	10.338	9	2.772	4	2
3	prometheus-community/helm-charts	39,7	6.169	5.382	70	1.799	193	84
4	helmfile/helmfile	20	5.176	359	31	394	25	11
5	open-compass/opencompass	18,7	7.313	842	30	183	376	92
6	komodorio/helm-dashboard	18,6	5.744	360	24	57	19	1
7	omerbsezer/Fast-Kubernetes	11,8	3.621	1.359	0	5	4	3
8	stackrox/kube-linter	10,9	3.499	273	10	94	89	12
9	helm/chartmuseum	9,9	3.837	407	1	88	116	25
10	roboll/helmfile	7,5	4.034	552	0	224	527	11

RAG / Vektorsuche

Rank	Repository	Score	Stars	Forks	Commits (30 T.)	Mitwirkende	Open issues	Open PRs
1	qdrant/qdrant	68,8	34.036	2.584	532	191	694	205
2	milvus-io/milvus	67,6	45.676	4.183	225	368	1.289	359
3	langchain4j/langchain4j	38,1	12.891	2.467	158	461	853	243
4	chroma-core/chroma	33,9	29.075	2.447	45	194	788	456
5	simdjson/simdjson	32,5	24.155	1.271	28	191	136	16
6	Tencent/APIJSON	27,9	18.398	2.275	31	96	241	3
7	datawhalechina/all-in-rag	26,9	10.422	5.166	2	11	23	8
8	svcvit/Awesome-Dify-Workflow	14	10.747	1.082	0	2	24	6
9	gka/chroma.js	13,2	10.575	551	0	43	76	15
10	std-microblock/chromatic	10	8.366	394	0	1	4	0

Snapshot-Daten, kein Live-GitHub-Stream. Updates laufen per geplanter REST/GraphQL-Erfassung, JSON-Schreiben und Rebuild.

GitHub‑zentrierte Signale

Repositories, Stars, Forks, Issues/PRs, Release‑Takt und Topics; Aggregation nach Organisation, Sprache oder Themen.

Projekte auf GitHub Explore ansehen

Kernmetriken

Star‑Velocity, Commit‑Frequenz, Issue/PR‑Aktivität, Beitragsverteilung, Release‑Rhythmus—mit Tages/Wochen/Monats‑Fenstern für Zeitreihen oder Heatmaps.

Ingestion und Aktualisierung

GitHub REST/GraphQL, öffentliche Datensätze oder Aggregatoren wie RSSHub; Actions zum periodischen Abruf und Normalisieren, Schreiben nach JSON oder D1, dann statische Seiten oder Read‑only‑Edge‑APIs.

Typen und Domänen

LLM / Inferenz‑Stack (Serving, Kernel, Quantisierung etc.)
Agenten, Multi‑Agenten, Tool‑Nutzung und Orchestrierung
Evals, Benchmarks, Datensätze und reproduzierbare Leaderboards
Training, Fine‑Tuning, Alignment und Engineering‑Toolchains
Lokales Deployment, Inferenz‑Services, Edge/Endpoint
Daten, RAG, Retrieval und Vector‑Store‑Ökosysteme

Häufige Fehldeutungen

Stars mit Modellqualität gleichsetzen—Hype kann vom Benchmark abweichen.
Nur kurze Zeitfenster—Spitzen können Kampagnen sein; längere Commit-Kurven tragen mehr Signal.
Issues/PRs ignorieren—gesunde Projekte zeigen oft dichte Wartung jenseits der Stars.
Persönliche und Organisations-Repos mischen—Governance und Release-Takt können stark abweichen.
GitHub als gesamtes Ökosystem sehen—wichtige Arbeit kann proprietär oder woanders gehostet sein.

Wöchentlicher KI-Branchenüberblick

Stand 18. August 2026 umfasst diese Hugging-Face-Hotlist 50 Einträge. Nach 24-Stunden-Wachstum fallen „Muse-Glimmer-30B-GGUF“, „LTX-2.5“, „Qwen3.8-27B-FP8“, „Qwen3.8-27B“, „Qwen3.8-27B-Uncensored-GGUF“ auf. Häufig oben vertretene Organisationen: Qwen, unsloth, MiniMaxAI, deepseek-ai—ein Hinweis auf schnelle Releases, Forks und Pipeline-Updates.

Etwa 48 Zeile(n) sind in diesem Snapshot als neu gelistet markiert—oft neue Gewichte, Quantisierungen oder Pipeline-Tags.

Tokens / Hosting: Hacker News hebt „Why is Anthropic's public writing style so unlike Claude's?“ hervor; Hacker News hebt „I wasted four months waiting for Anthropic to review my MCP“ hervor; Hacker News hebt „OpenAI has the right strategy, but is unlucky“ hervor. Hinweise zu Preisen, Regionen und Rollout-Tempo für gehostete Inferenz.

Der gruppierte GitHub-Vergleich umfasst 8 Ökosystem-Segmente, um Modell-Hype mit Engineering-Aktivität (Stars, Commits, Issues/PRs) zu vergleichen.

Zur Startseite

Angesagte Modelle

Qwen3.8-27B 🚀 New

Qwen3.8-27B-GGUF New

Qwen3.8-2.4T-A95B 🚀 New

LTX-2.5 🚀 New Hot

MiniMax-Music3 🚀 New

Muse-Glimmer-30B 🚀 New

MiniMax-H3 New

DeepSeek-V4-Pro-0813 🚀 New

Qwen3.8-27B-FP8 🚀 New Hot

Qwen3.8-27B-Uncensored-FP8 🚀 New

DeepSeek-V4-Flash-0731 New

Qwen3.8-27B-Uncensored-GGUF 🚀 New

Minimax-h3-Turbo 🚀 New

NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4 🚀 New

Kimi-K3 New

Qwen3.8-27B-NVFP4 🚀 New

Muse-Glimmer-30B-GGUF 🚀 New Hot

Anima-2.9B 🚀 New

Ling-3.0-tiny 🚀 New

Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-NEO-MAX-MTP-GGUF New

Qwen3.8-2.4T-A95B-FP8 🚀 New

MiniMax-H3 New

dots3-note-prev 🚀 New

Qwen-Fixed-Chat-Templates New

MiniMax-Music-3 🚀 New

NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16 🚀 New

LFM2.5-VL-3B 🚀 New

MiniMax-H3-Realism-People-LoRA 🚀 New

10Eros-Max New

MiniMax-H3-Turbo-Lora New

LFM2.5-2.6B 🚀 New

Qwen3.8-27B-ABLITERATED-GGUF 🚀 New

Qwen3.8-27B-Ridge-GGUF 🚀 New

needle2 🚀 New

Qwen3.6-27B-Uncensored-HauhauCS-Aggressive 🚀

Qwen3.8-27B-AEON-ULTIMATE-UNCENSORED-BF16 🚀 New

IndexTTS-2.5 🚀 New

Muse-Glimmer-30B-GGUF 🚀 New Hot

Qwen3.8-27B-Heretic-Abliterated-Uncensored-GGUF 🚀 New

North-Micro-Vision-Instruct 🚀 New

Qwen3.8-9B 🚀 New

Qwen3.8-2.4T-A95B-GGUF 🚀 New

Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-MTP-GGUF

Qwen3.8-27B-GGUF 🚀 New

Qwen3.8-27B-Cold-Fusion-GAIN-V1.1 New

MiniMax-H3_comfy New

MiniMax-H3-GGUF 🚀 New

Qwen3.8-27B-Abliterated-MLX New

Qwen3.8-27B-Uncensored-MLX New

Huihui-Qwen3.8-27B-abliterated 🚀 New

Gruppierter Repository-Vergleich

Agent

Multi-Agent

Inferenz

LLM-Bibliothek

Trainings-Framework

Serving

Evaluation

RAG / Vektorsuche

GitHub

GitHub‑zentrierte Signale

Kernmetriken

Ingestion und Aktualisierung

Typen und Domänen

Häufige Fehldeutungen

Weiterlesen

Wöchentlicher KI-Branchenüberblick

Qwen3.8-27B New

Qwen3.8-2.4T-A95B New

LTX-2.5 New Hot

MiniMax-Music3 New

Muse-Glimmer-30B New

DeepSeek-V4-Pro-0813 New

Qwen3.8-27B-FP8 New Hot

Qwen3.8-27B-Uncensored-FP8 New

Qwen3.8-27B-Uncensored-GGUF New

Minimax-h3-Turbo New

NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4 New

Qwen3.8-27B-NVFP4 New

Muse-Glimmer-30B-GGUF New Hot

Anima-2.9B New

Ling-3.0-tiny New

Qwen3.8-2.4T-A95B-FP8 New

dots3-note-prev New

MiniMax-Music-3 New

NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16 New

LFM2.5-VL-3B New

MiniMax-H3-Realism-People-LoRA New

LFM2.5-2.6B New

Qwen3.8-27B-ABLITERATED-GGUF New

Qwen3.8-27B-Ridge-GGUF New

needle2 New

Qwen3.6-27B-Uncensored-HauhauCS-Aggressive

Qwen3.8-27B-AEON-ULTIMATE-UNCENSORED-BF16 New

IndexTTS-2.5 New

Muse-Glimmer-30B-GGUF New Hot

Qwen3.8-27B-Heretic-Abliterated-Uncensored-GGUF New

North-Micro-Vision-Instruct New

Qwen3.8-9B New

Qwen3.8-2.4T-A95B-GGUF New

Qwen3.8-27B-GGUF New

MiniMax-H3-GGUF New

Huihui-Qwen3.8-27B-abliterated New