Top 20 AI GitHub Repositories of 2026

1. 1. OpenClaw leads on combined influence and activity

   openclaw/openclaw ranks #1 with a DataAIHub Score of 95.21. In this Top 20 it also leads stars, forks, and recent commit volume — showing that the top overall score can align with multiple strong signals, not stars alone.

2. 2. Agent and orchestration stacks are heavily represented

   About half of the Top 20 are agent frameworks, LLM app platforms, or orchestration layers — including hermes-agent, Dify, LangChain, LiteLLM, AutoGPT, LobeHub, Langflow, and RAGFlow. The mid-year list is not limited to model libraries.

3. 3. Inference and local serving remain core infrastructure

   llama.cpp (#6), vLLM (#8), and Ollama (#18) all appear in the Top 20, underscoring sustained community weight around open inference and local model serving alongside training stacks such as PyTorch and Transformers.

4. 4. Star popularity is not the same as ranking position

   affaan-m/ECC has the second-highest star count in the Top 20 (237.7K) but ranks #17. Repositories with fewer stars can place higher when forks, contributors, recent commits, and releases are stronger under the cohort-relative DataAIHub Score.

5. 5. High-frequency release projects stand out

   llama.cpp, LobeHub, LiteLLM, and LangChain show unusually high release totals in this snapshot. Releases are only 5% of the DataAIHub Score, but they can still contribute meaningfully when combined with other strong repository signals.

Repository names link to the official GitHub repository. Scores are DataAIHub Scores (0–100) relative to the 326-repository eligible cohort for this snapshot — not relative to the Top 20 alone.

Individual signals often have different leaders than the overall ranking. The repository with the most stars is not necessarily #1 on the DataAIHub Score.

Profiles below use verified GitHub metadata and the same snapshot metrics used for scoring. Ranking signals summarize the most notable measurable signals for each repository in this Top 20.

The DataAIHub Score combines multiple repository signals. Each metric is log-normalized relative to the maximum value in the eligible ranking cohort, then combined with fixed weights and scaled to 0–100.

Log normalization reduces the effect of extreme scale differences — a repository with millions of stars would otherwise dominate raw arithmetic — while still rewarding stronger signals.

Scores are relative to the eligible repository cohort used for this ranking. A score is not an absolute universal GitHub quality score. It represents performance relative to the repositories included in this DataAIHub ranking.

Repository popularity does not necessarily mean technical superiority. This ranking is data-driven and based on measurable repository signals.

• Stars — Community interest and adoption signal on GitHub.
  35%
• Forks — Developer usage and extension signal.
  20%
• Commit activity — Recent development activity (last ~30 days via commit-activity stats).
  20%
• Contributors — Breadth of project participation (pipeline capped at 1,000).
  10%
• Watchers — Ongoing notification interest (GitHub subscribers_count when available).
  10%
• Releases — Published release activity on GitHub.
  5%

For the live daily series and additional detail, see the GitHub rankings methodology.

• GitHub stars measure attention, not software quality.
• Large mature projects naturally accumulate stronger historical signals.
• Commit volume does not necessarily equal meaningful development.
• Some projects conduct development across multiple repositories.
• GitHub metrics cannot capture enterprise or private adoption.
• Ranking results depend on the eligible repository cohort.
• This is a snapshot; GitHub activity changes continuously.

• Live AI GitHub Rankings →
• GitHub Ranking Methodology →
• GitHub Trending →
• Guides →
• AI Tools Rankings →