Back to Gap Map Model components / Embeddings & retrieval models

SigLIP

open weights / Overall score: 4.0(strong)

SigLIP replaces CLIP's softmax contrastive loss with a pairwise sigmoid loss, which needs no global view of the batch and so trains well at both very large and quite small batch sizes. SigLIP 2 folds captioning pretraining, self-distillation, masked prediction and online data curation into the same recipe and adds multilingual coverage and native-aspect-ratio variants. The checkpoints serve two audiences at once: image-text retrieval in their own right, and the default frozen vision tower for a large share of current open VLMs.

The Hub cards for SigLIP 1 were written by Hugging Face rather than by the release team, and both generations put their evaluation tables in images; the numbers used here come from big_vision's own SigLIP 2 checkpoint table instead.

Openness

3 high confidence

3.0

weights: open(44 ungated Apache-2.0 google/siglip* repos on the Hub, plus the .npz checkpoints served directly from the big_vision bucket)
data: closed(WebLI, a private Google web corpus
code: partial(big_vision ships a SigLIP trainer and a LiT/COCO demo config, but records 'code TODO' against the SigLIP paper entry and publishes no WebLI pretraining config)
license: Apache-2.0(OSI)

Apache-2.0 weights with no gate, which is more permissive than most of Google's model releases, over a corpus that is named and nothing more. big_vision gets closer to a pipeline than CLIP's repository does - there is a real SigLIP trainer in it - but the entry for the SigLIP paper says 'code TODO' and the config that would reproduce a WebLI run is not there, so this is partial rather than open.

https://huggingface.co/google/siglip-so400m-patch14-384/raw/main/README.md recorded 2026-10-20
`license: apache-2.0`; 'SigLIP is pre-trained on the WebLI dataset'; 'first released in this repository' pointing at google-research/big_vision.
https://huggingface.co/google/siglip2-base-patch16-224/raw/main/README.md recorded 2026-10-20
`license: apache-2.0`; 'SigLIP 2 is pre-trained on the WebLI dataset (Chen et al., 2023)'; evaluation results supplied as an image.
https://raw.githubusercontent.com/google-research/big_vision/main/README.md recorded 2026-10-20
the SigLIP paper entry reads 'Resources: colab and models, code TODO'; an 'Image-text training with SigLIP' section launches big_vision.trainers.proj.image_text.siglip with the siglip_lit_coco.py config.

Adoption

4 high confidence

4.0

Adoption is measured as Hugging Face downloads summed across every SigLIP and SigLIP 2 checkpoint Google publishes under google/siglip*, declared as a family in both the Transformers and JAX formats and led by the SigLIP 2 giant-opt and base checkpoints. The timm re-hosts of the same checkpoints are excluded as mirrors.

https://huggingface.co/api/models?author=google&search=siglip&limit=100&sort=downloads&direction=-1 recorded 2026-10-17
44 google siglip repos, all declared, with `downloads` (trailing 30 days) summing to 9,952,373, led by siglip2-giant-opt-patch16-384 at 1,737,175; the 17 *-jax format repos draw none.

Capability

4 high confidence

4.0

SigLIP is current-generation and multilingual, posting zero-shot ImageNet accuracy about five points ahead of the best OpenCLIP checkpoint, and it serves as the default vision tower for much of the open VLM stack. It falls short of the top spot in the open image-text field, though the gap is narrow rather than absent.

https://huggingface.co/google/siglip2-base-patch16-224/raw/main/README.md recorded 2026-09-11
'SigLIP 2 models outperform their SigLIP counterparts at all model scales in core capabilities, including zero-shot classification, image-text retrieval, and transfer performance'.
https://raw.githubusercontent.com/google-research/big_vision/main/big_vision/configs/proj/image_text/README_siglip2.md recorded 2026-09-11
checkpoint table, INet 0-shot / COCO T-I / COCO I-T: g-opt/16 384 -> 85.0 / 56.1 / 72.8; So400m/16 512 -> 84.3 / 56.0 / 71.3; B/16 224 -> 78.2 / 52.1 / 68.9.

Verified 2026-09-11