AI Hippo

Last updated:

Hottest open-source projects in the last 24 hours

sindresorhus/awesome · ⭐ 453,805 freeCodeCamp/freeCodeCamp · ⭐ 442,475 EbookFoundation/free-programming-books · ⭐ 385,282 openclaw/openclaw · ⭐ 353,817 kamranahmedse/developer-roadmap · ⭐ 352,641 vinta/awesome-python · ⭐ 291,625 awesome-selfhosted/awesome-selfhosted · ⭐ 285,196 facebook/react · ⭐ 244,482 sindresorhus/awesome · ⭐ 453,805 freeCodeCamp/freeCodeCamp · ⭐ 442,475 EbookFoundation/free-programming-books · ⭐ 385,282 openclaw/openclaw · ⭐ 353,817 kamranahmedse/developer-roadmap · ⭐ 352,641 vinta/awesome-python · ⭐ 291,625 awesome-selfhosted/awesome-selfhosted · ⭐ 285,196 facebook/react · ⭐ 244,482

Open source · Live · GitHub + HF

Tendências OSS no GitHub

Além dos rankings, observamos mudanças estruturais no ecossistema open source no GitHub: atividade, atenção e calor comunitário por domínio e tipo de projeto.

Snapshot

Hot models

24h Δ

Sorted by estimated VRAM (GB), ascending; unknown estimates go last.

Grouped repository comparison

Each section lists repositories of the same class (e.g., Agent: OpenClaw, NanoBot, PicoClaw). Column meanings are consistent across groups. “Commits (30d)” follows the scraper’s rolling window.

Snapshot:

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 (30d) Contributors Open issues Open PRs
1 openclaw/openclaw 82,6 353.817 71.480 920 2.100 18.075 380
2 langchain-ai/langchain 50,4 133.041 21.954 412 4.100 513 98
3 Significant-Gravitas/AutoGPT 37,7 183.295 46.227 42 1.200 391 52
4 All-Hands-AI/OpenHands 19,2 70.963 8.899 132 460 390 72
5 crewAIInc/crewAI 18,4 48.516 6.619 124 890 518 36
6 langchain-ai/langgraph 16,9 28.878 4.939 198 420 490 45
7 hkuds/nanobot 15,8 38.909 6.785 156 180 936 42
8 qwibitai/nanoclaw 14,2 27.022 11.466 112 95 623 28
9 sipeed/picoclaw 11,9 27.963 3.963 88 64 283 18
10 Pythagora-io/gpt-pilot 10,8 33.777 3.502 36 130 243 24

Multi-agent

Rank Repository Score Stars Forks Commits (30d) Contributors Open issues Open PRs
1 browser-use/browser-use 84 86.945 10.033 210 210 236 48
2 geekan/MetaGPT 63,5 66.886 8.474 78 390 113 30
3 microsoft/autogen 62,5 56.925 8.558 86 512 756 34
4 OpenInterpreter/open-interpreter 45 63.065 5.447 38 210 313 28
5 SWE-agent/SWE-agent 27,2 18.959 2.051 64 120 36 22
6 ag2ai/ag2 24,3 4.380 580 74 240 223 26
7 langroid/langroid 17,5 3.959 368 52 85 64 12
8 microsoft/JARVIS 16,5 24.624 2.098 10 72 1.312 5
9 microsoft/TaskWeaver 16,2 6.139 771 28 96 53 8
10 aorwall/moatless-tools 10,7 637 60 18 68 20 9

Inference

Rank Repository Score Stars Forks Commits (30d) Contributors Open issues Open PRs
1 vllm-project/vllm 80,7 76.005 15.415 380 920 4.153 180
2 ggerganov/llama.cpp 71 102.901 16.637 142 684 1.441 62
3 microsoft/onnxruntime 38 19.818 3.819 165 890 1.349 120
4 sgl-project/sglang 31 25.614 5.268 210 210 2.754 95
5 apache/tvm 29,1 13.264 3.850 88 760 175 85
6 Tencent/ncnn 20,6 23.071 4.414 48 320 1.184 42
7 mlc-ai/mlc-llm 18,6 22.414 1.998 58 210 300 21
8 pytorch/executorch 16,3 4.493 929 136 190 1.894 36
9 NVIDIA/TensorRT-LLM 15,5 13.331 2.271 72 210 1.266 42
10 internlm/lmdeploy 13,6 7.770 681 86 150 586 30

LLM library

Rank Repository Score Stars Forks Commits (30d) Contributors Open issues Open PRs
1 huggingface/transformers 79,3 159.149 32.820 198 2.950 2.360 156
2 google-gemini/gemini-cli 59 100.792 13.003 320 210 2.950 48
3 BerriAI/litellm 36,5 42.817 7.133 220 260 2.390 52
4 unslothai/unsloth 34,7 60.803 5.232 165 85 1.120 62
5 ml-explore/mlx 17,3 25.279 1.672 42 120 162 28
6 QwenLM/Qwen 16,9 20.957 1.781 40 110 34 15
7 OpenBMB/MiniCPM 16,2 8.786 562 62 130 19 16
8 Lightning-AI/litgpt 13,7 13.291 1.418 26 170 249 20
9 mistralai/mistral-inference 11,2 10.762 1.035 8 42 171 8
10 LLaVA-VL/LLaVA-NeXT 10,5 4.628 456 18 90 361 12

Training framework

Rank Repository Score Stars Forks Commits (30d) Contributors Open issues Open PRs
1 tensorflow/tensorflow 73,4 194.623 75.264 210 4.200 4.301 420
2 pytorch/pytorch 65,3 98.998 27.457 620 5.120 18.278 820
3 keras-team/keras 30,9 63.956 19.739 94 1.500 268 96
4 jax-ml/jax 24,8 35.351 3.514 180 890 2.205 120
5 tinygrad/tinygrad 23,4 32.290 4.047 170 460 183 74
6 PaddlePaddle/Paddle 21,8 23.809 5.983 140 980 1.507 150
7 apache/mxnet 15,4 20.810 6.715 12 890 2.007 42
8 mindspore-ai/mindspore 14,9 4.679 756 88 620 220 63
9 facebookresearch/xformers 14,5 10.415 774 84 240 377 55
10 openxla/xla 13,8 4.148 777 92 210 634 44

Serving

Rank Repository Score Stars Forks Commits (30d) Contributors Open issues Open PRs
1 ollama/ollama 69,3 168.428 15.500 124 420 2.893 72
2 ray-project/ray 58,7 42.059 7.421 210 2.300 3.587 420
3 mudler/LocalAI 33,3 45.181 3.902 95 210 174 48
4 skypilot-org/skypilot 21,7 9.801 1.026 84 125 374 24
5 lm-sys/FastChat 20,3 39.444 4.795 16 240 1.015 18
6 huggingface/text-generation-inference 19,3 10.830 1.261 64 180 324 36
7 KServe/KServe 18,9 5.317 1.431 66 320 671 62
8 vllm-project/aibrix 18,2 4.714 547 72 60 320 10
9 triton-inference-server/server 15,3 10.533 1.751 42 120 872 28
10 bentoml/OpenLLM 12,8 12.273 806 22 105 12 14

Evaluation

Rank Repository Score Stars Forks Commits (30d) Contributors Open issues Open PRs
1 explodinggradients/ragas 69,4 13.312 1.348 118 180 352 32
2 confident-ai/deepeval 66,7 14.676 1.342 96 160 260 20
3 openai/evals 65,2 18.165 2.915 14 310 183 24
4 EleutherAI/lm-evaluation-harness 59,6 12.099 3.176 48 320 794 42
5 open-compass/opencompass 34,6 6.847 755 52 120 444 18
6 RUC-NLPIR/FlashRAG 24,9 3.446 298 40 88 34 11
7 huggingface/lighteval 18,9 2.372 451 38 42 301 12
8 stanford-crfm/helm 16,1 2.740 370 12 85 70 8
9 lmarena/arena-hard-auto 13,2 1.015 150 20 55 15 7
10 prometheus-eval/prometheus-eval 10,7 1.066 70 14 42 14 6

RAG / vector search

Rank Repository Score Stars Forks Commits (30d) Contributors Open issues Open PRs
1 langgenius/dify 89,4 137.074 21.437 260 980 899 160
2 run-llama/llama_index 48,9 48.472 7.185 165 890 271 85
3 FlowiseAI/Flowise 42,6 51.742 24.096 96 520 801 84
4 Mintplex-Labs/anything-llm 35,7 58.014 6.272 120 340 320 55
5 milvus-io/milvus 27 43.716 3.948 120 420 1.158 95
6 deepset-ai/haystack 21,8 24.790 2.704 72 420 117 28
7 qdrant/qdrant 20,1 30.220 2.161 95 210 511 42
8 vercel/ai-chatbot 16,8 20.067 6.468 54 160 14 22
9 weaviate/weaviate 11,3 15.982 1.248 64 180 544 36
10 chroma-core/chroma 10,8 27.338 2.183 42 120 557 28

Demo data—not live GitHub. Production: schedule REST/GraphQL ingestion, write JSON, then rebuild.

Métricas de exemplo

Velocidade de estrelas, cadência de commits recentes, atividade em issues/PRs, distribuição de contribuidores, ritmo de releases—por dia/semana/mês para séries ou mapas de calor.

Ingestão e atualização

GitHub REST/GraphQL, datasets públicos ou agregadores como RSSHub; Actions programados para normalizar e gravar JSON ou D1, depois páginas estáticas ou APIs somente leitura no edge.

Tipos e domínios que observamos

  • LLM / pilha de inferência (serving, kernels, quantização, etc.)
  • Agentes, multi‑agente, uso de ferramentas e orquestração
  • Avaliações, benchmarks, datasets e tabelas reproduzíveis
  • Treinamento, fine‑tuning, alinhamento e toolchains de engenharia
  • Implantação local, serviços de inferência e edge/endpoint
  • Dados, RAG, recuperação e ecossistemas de vector stores

Common misreadings

  • Equating stars with model quality—hype may diverge from benchmark strength.
  • Relying on short windows only—spikes may reflect campaigns; longer commit curves carry more signal.
  • Ignoring issues/PRs—healthy projects often show dense maintenance beyond stars.
  • Mixing personal and organization repositories—governance and release cadence may differ materially.
  • Treating GitHub as the entire ecosystem—important work may be closed-source or hosted elsewhere.

Mais leitura

Weekly AI industry brief (auto-generated)

For the week ending Apr 10, 2026, this Hugging Face hot list tracks 8 entries. By 24-hour growth, standouts include "gemma-4-E4B-it", "Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distil…", "OmniVoice", "Gemma-4-31B-JANG_4M-CRACK", "GLM-5.1". Organizations appearing most often near the top include google, dealignai, zai-org, netflix—a sign that releases, forks, and task pipelines remain fast-moving.

About 1 row(s) are flagged as newly listed in this snapshot, which usually means fresh weights, quant formats, or pipeline tags entering the cohort.

On infrastructure: Hacker News highlights “Potential Anthropic Price Increases”; Hacker News highlights “Anthropic may have leaked Claude Code source on purpose”; Hacker News highlights “LLM scraper bots are overloading acme.com's HTTPS server”. These cues tie to pricing, regions, and rollout cadence for hosted inference.

The grouped GitHub comparison below spans 8 ecosystem segments so you can contrast model buzz with engineering activity (stars, commits, issues/PRs). This text is auto-generated from the same snapshot as the tables and is for research context only—not financial or legal advice.

Voltar ao início