中文

Benchmark Radar

I kept running into new benchmarks while doing benchmark research, so I built a crawler that continuously collects benchmark-related signals from across the web. It pulls evidence from 37 public sources every day, and keeps updating.

Find a benchmark in seconds, then see how model scores change over time. Click the GIF below to watch SWE-bench Verified move toward saturation.

See the dashboard

Today: everything that showed up in the last 24 hours, scored and ranked, plus a short daily briefing that says what changed and links the evidence it used.

Leaderboard: which benchmarks labs actually report in their model cards, and how scores on each one climb until there is almost no headroom left.

Use it

• Open the dashboard — today's insights, trends, popular benchmarks, model-card adoption, and more
• Query it locally — install and use the offline CLI
• Subscribe via RSS — get new benchmark signals every day
• Download the complete dataset — free, public, machine-readable JSON; no crawler or contact required
• Contribute — add benchmarks, model cards, sources, or fixes

If Benchmark Radar saves you research time, star the repository. It helps other eval builders find it.

Query it locally (CLI version)

npx skills add ktwu01/benchmark-radar

Then ask your coding agent about benchmarks. It installs the command-line tool and downloads the data to your computer the first time you ask. What it does is written in the setup and usage guide.

More

• Design principles
• Scoring rubric
• Model-card adoption data
• Public corpus schema
• Citation information
• Technical report
• Configuration
• Developer setup: python -m pip install -e '.[dev]' && benchmark-radar normalize-external
• Support / bugs: open an issue
• Contact: @ktwu01

Licensing

Software: MIT License.

Technical report and original editorial content: CC BY-NC 4.0. Commercial republication, resale, paid newsletters, dataset packaging, or commercial product integration requires prior written permission from Koutian Wu. Third-party source material remains under its original terms.

Join the WeChat group

Scan the QR code to join the WeChat group for daily benchmark updates and eval discussions:

Contributors

Thanks to everyone who helps make Benchmark Radar more useful.

Citation

If Benchmark Radar supports your research or evaluation work, please cite the technical report:

@misc{wu_2026_22167102,
  author       = {Wu, Koutian and Zhou, Junjie},
  title        = {Benchmark Radar v0.9.0: Technical Report},
  month        = aug,
  year         = {2026},
  publisher    = {Zenodo},
  version      = {0.9.0},
  doi          = {10.5281/zenodo.22167102},
  url          = {https://doi.org/10.5281/zenodo.22167102}
}

See CITATION.cff for machine-readable citation metadata.

Star History

Acknowledgements

The daily evidence feed is built on public data from arXiv, GitHub Search, GitHub organizations, GitHub Releases, Hugging Face datasets and Spaces, Hugging Face Papers, Crossref, OpenAlex, OpenReview, Kaggle datasets, Zenodo, Semantic Scholar, Brave Search, and Hacker News, plus first-party lab feeds from OpenAI, Google AI, Google DeepMind, Google Research, Meta Research, Microsoft Research, AWS Machine Learning, Apple Machine Learning Research, NVIDIA AI Blog, NVIDIA Developer, Hugging Face Blog, Ai2, Mistral AI, Together AI, Sakana AI, Qwen, Ollama, Stability AI, Nomic AI, Replicate, IBM Research, Databricks, LangChain, and Meituan Engineering.

The frontier-model score layer, including the SWE-bench Verified timeline above, is built on benchmark data collected by LLM Stats. Thank you for keeping that data open.

A special thank you to Xiaopai Liu (@bourneliu66) for the shout-out on X, and to his daily builder brief, BuilderPulse.

Internal documentation

• SEO and indexing guide
• Benchmark logo gallery

contribute score

See the public contribution-score ledger and rules.