EVA-01
Open in the interactive archive →
Download the covers: Floppy card · VHS cover · Vinyl cover
A to Z
- A · App
- No
- B · Built By
- Beijing Academy of AI
- C · CEO & Politics
- N/A (research institute leadership) — Not publicly documented (PRC-based; operates under Chinese regulatory environment)
- D · Date
- 2023
- E · Expertise
- Intermediate+
- F · Free Tier
- Yes
- G · Generates
- Text only
- H · Deployment
- Local (open weights)
- I · Inputs
- Text
- J · Jurisdiction
- China
- K · Key Staffers
- BAAI (Chinese gov-funded research)
- L · Languages
- Chinese-primary
- M · Model Type
- Vision/CV
- N · Safety Issues
- Computer-vision risks: dataset bias, domain shift, privacy concerns in images, and unreliable performance outside benchmark-like settings.
- O · Open Source
- Yes — open weights
- P · Pricing
- Free (open weights, MIT)
- Q · AIDb Rating
- 4/10 · AIDb v2.0 (rule-based): Capability 4/10 · Currency 4/10 · Access 2/10 · Significance 4/10 → 3.70
- R · What Makes It Special
- EVA-01, the EVA vision foundation model: a plain Vision Transformer scaled to one billion parameters with only publicly accessible data, pretrained to reconstruct masked-out, image-text aligned vision features from visible image patches. Its authors reported new records in image recognition, video action recognition, object detection and segmentation without heavy supervised training, and found that starting a giant CLIP's vision tower from EVA stabilizes its training. Code and models are open.
- S · Stock / Status
- Government-funded
- T · Training Cutoff
- [UNKNOWN]
- U · Users
- [UNKNOWN]
- V · Voice Chat
- No
- W · Weights
- Downloadable
- X · X Handle
- @BAAIBeijing
- Y · Year Founded
- [UNKNOWN]
- Z · Bottom Line
- Open-weights Vision/CV from Beijing Academy of AI, released 2023.
Data completeness 88% · Last verified 2026-06-11 · Primary source ↗
Cite this
“EVA-01.” AIDb — The AI Database. Verified 2026-06-11. https://theaidatabase.com/m/eva-01-beijing-academy-of-ai.html