This is the public data package for BioXArena: 9 domains, 76 tasks. Each task folder contains public/description.md, public/samplesubmission.csv, the training and test inputs, and any modality-specific assets. We host the BioXArena benchmark on CodaBench.
Dataset Card
This is the public data package for BioXArena: 9 domains, 76 tasks. Each task folder contains public/description.md, public/samplesubmission.csv, the training and test inputs, and any modality-specific assets. We host the BioXArena benchmark on CodaBench. Welcome to evaluate your coding agents and climb the leaderboard. This release replaces all earlier versions of the data. Since the previous release, 55 of the 76 tasks were rebuilt or corrected to remove label leakage, shortcuts that let a model score well without solving the task, synthetic or misaligned labels, and input defects; several tasks were re-split (for example scaffold-, study- or site-level hold-out) and a few evaluation…
Excerpt from the card by Mbzuai Ai4bio, licensed cc-by-nc-4.0.
Details
- Repository
- mbzuai-ai4bio/BioXArena-Data-Public
- Publisher
- Mbzuai Ai4bio
- Task category
- Not stated by the source
- Tags
- Not stated by the source
- Size category
- Not stated by the source
- Languages
- Not stated by the source
- Revision
- 3787cf1f6eb217d09a109e0baca0cb7e15298b30
- Last updated
- 2026-10-15
Files
13 files, 97.8 GB in total.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| BioXArena-Data-Public.tar.gz | Data | 48.9 GB | cd9674d3820d |
| Data-Public-by-Domain/MANIFEST.txt | Data | 5.2 KB | — |
| Data-Public-by-Domain/chemical-biology-8tasks-all.tar.gz | Data | 23.7 MB | 4f171714af78 |
| Data-Public-by-Domain/imaging-8tasks-all.tar.gz | Data | 24.9 GB | 9b8cb1ae9b51 |
| Data-Public-by-Domain/network-biology-8tasks-all.tar.gz | Data | 27.3 MB | 25680d05dbe2 |
| Data-Public-by-Domain/perturbation-dynamics-8tasks-all.tar.gz | Data | 272.4 MB | 42ba3ff879eb |
| Data-Public-by-Domain/phenotype-disease-8tasks-all.tar.gz | Data | 3.2 GB | 34fee19ad95b |
| Data-Public-by-Domain/sequence-10tasks-all.tar.gz | Data | 176.8 MB | 1ba0016f280c |
| Data-Public-by-Domain/single-cell-10tasks-all.tar.gz | Data | 4.0 GB | b50bba8b39a7 |
| Data-Public-by-Domain/structure-8tasks-all.tar.gz | Data | 8.5 GB | d07b3f26cc7f |
| Data-Public-by-Domain/text-integrated-8tasks-all.tar.gz | Data | 7.7 GB | 8ee3ade93eac |
| README.md | Documentation | 24.6 KB | — |
| .gitattributes | Repository | 2.5 KB | — |
License and Download
- License
- cc-by-nc-4.0
- Access
- No access gate
Released by Mbzuai Ai4bio through its official repository on Hugging Face. Read the license.