Hugging Face is displaying Every Eval Ever evaluation results on its model pages, according to reporting published by the organization on June 30, 2026. TechCrunch independently confirmed the presentation of the benchmark scores across model pages on the platform. The update puts evaluation metrics directly into the core documentation areas where open models are hosted.
Original reporting from Hugging Face on June 30, 2026, outlined the integration of Every Eval Ever data into standard model listings. TechCrunch separately reported the publication of these benchmark results on the platform. Neither publication stated whether individual model creators must manually request the inclusion of Every Eval Ever scores on their pages.
TechCrunch did not clarify whether the Every Eval Ever metrics cover all model categories hosted by Hugging Face or focus on specific model types. Hugging Face gave no details on how frequently evaluation scores are updated on model pages. The organization also did not say whether users can filter model search results using Every Eval Ever benchmark data.
Hugging Face did not publish specific administrative limits or submission guidelines regarding how Every Eval Ever results are processed for new model uploads. TechCrunch provided no details regarding potential API access changes linked to the displayed evaluation data. The reporting did not indicate whether third-party benchmark providers can contribute data to the model pages.
Model documentation on Hugging Face reflects the Every Eval Ever integration as reported on June 30, 2026. Hugging Face did not disclose plans for bringing other evaluation datasets to its model pages in subsequent updates. TechCrunch gave no indication of whether hardware requirements or compute constraints affect how these evaluation scores are displayed.
