- Released
- 3y ago
- Size
- —parameters
CLIP image-text embedding model trained on DFN-5B, a dataset filtered by Apple's Data Filtering Networks.
How it compares
Frequently asked questions
Privacy and data
Yes. DFN5B CLIP ViT-H/14 is published with open weights, so it can be self-hosted for full control of your data.
Capabilities and specs
It accepts text and image inputs only. Audio, video and PDF inputs are not supported.
Quality and benchmarks
Matched on model kind and release date, with no comparable price, the closest alternatives are bge-large-en-v1.5 by BAAI at a blended price of $0.0075 per 1M tokens, bge-base-en-v1.5 by BAAI at $0.00375 per 1M tokens, and GTE-Base by Thenlper at $0.00375 per 1M tokens.
bge-large-en-v1.5DFN5B CLIP ViT-H/14 vs bge-large-en-v1.5bge-base-en-v1.5DFN5B CLIP ViT-H/14 vs bge-base-en-v1.5GTE-BaseDFN5B CLIP ViT-H/14 vs GTE-Base
Licensing and access
Open weights. DFN5B CLIP ViT-H/14 is released with open weights under the apple-amlr license.
The weights are available from the Hugging Face repository apple/DFN5B-CLIP-ViT-H-14-378.
Basics
DFN5B CLIP ViT-H/14 was made by Apple and released on October 30, 2023.
Yes. DFN5B CLIP ViT-H/14 is published with open weights.
Generated from Vertical Tabs data, updated November 2, 2026.





Discussion
No comments yet
Be the first to share your thoughts.