SAVRN Model Hub · Comparisons
clip-vit-large-patch14 vs siglip2-base-patch16-naflex
Clip-vit-large-patch14 has 428M parameters and siglip2-base-patch16-naflex has 375M parameters; at 16-bit, clip-vit-large-patch14 needs about 1 GB (1x MI300X from $1.85 an hour) and siglip2-base-patch16-naflex about 0.9 GB (1x MI300X from $1.85 an hour).
| Field | clip-vit-large-patch14 openai/clip-vit-large-patch14 | siglip2-base-patch16-naflex google/siglip2-base-patch16-naflex |
|---|---|---|
| Publisher | OpenAI | |
| Task | Zero shot image classification | Zero shot image classification |
| Modality | Other | Other |
| Parameters, as reported | 428M parameters | 375M parameters |
| Architecture | CLIPModel | Siglip2Model |
| Library | transformers | transformers |
| Context length | 77 tokens | Not stated |
| Repository size | 6.8 GB | 1.5 GB |
| Artifact formats | safetensors, pytorch, jax, tf | safetensors |
| License | Not stated | apache-2.0 |
| Access | Open weights, no gate | Open weights, no gate |
| Memory at 16-bit (weights and margin) | 1 GB | 0.9 GB |
| Cheapest GPUs at 16-bit, per hour | 1x MI300X, $1.85 | 1x MI300X, $1.85 |
| Memory at 4-bit (weights and margin) | 0.3 GB | 0.2 GB |
| Cheapest GPUs at 4-bit, per hour | 1x MI300X, $1.85 | 1x MI300X, $1.85 |
| Revision viewed | 32bd64288804 | b53b807d3a2d |
| Downloads reported by the hub | 8.3M | 953.8k |
| Last observed | 2026-09-19 | 2026-09-18 |
An evaluation row appears only where at least two of these models report the same benchmark with the same stated configuration, metric, unit and setup. Different evaluators stay named in each cell. Values are shown as reported: no unit conversion, no ranking.
SAVRN's Notes on clip-vit-large-patch14
The text side of this model stops at 77 tokens, which tells you what it is for: scoring images against short text labels, the zero-shot classification OpenAI built it to study, not reading documents. On disk it is 6.85 GB across 13 files, float32 in four formats, but loaded at 16-bit it needs 1.0 GB and at 4-bit 0.3 GB. The cheapest setup we list, one 192 GB MI300X at $1.85 per hour on-demand, is a card you would share; the question is throughput, not fit.
Our record carries no license for it; that is the first check: get the terms from the publisher's files before commercial use. The second is the card's warning that it was not developed for general deployment; study it in your context, and if faces are in your images, start with the FairFace paper on race, gender and age the record ties it to.
SAVRN's Notes on siglip2-base-patch16-naflex
Google built this encoder for three jobs: sorting images against text labels with no task-specific training, matching images to captions for retrieval, and serving as the vision tower inside a larger vision-language model. At 375 million parameters it needs 0.9 GB of memory at 16-bit precision, so hardware choice is about what else is on the card. The cheapest card in our table, one MI300X with 192 GB at $1.85 an hour on demand, carries it alongside the language model it feeds.
Apache 2.0 clears commercial use, modification and redistribution with the notices kept. Three papers describe the method, arXiv 2502.14786, 2303.15343 and 2209.06794, and the first titles it a multilingual vision-language encoder, which matters if your labels are not in English. The release is February 2025, so confirm your inference stack supports the siglip2 model type before ordering hardware.
Questions
Which is larger, clip-vit-large-patch14 or siglip2-base-patch16-naflex?
clip-vit-large-patch14 (428M parameters) is larger than siglip2-base-patch16-naflex (375M parameters), by the parameter counts their publishers report.
Which is cheaper to run, clip-vit-large-patch14 or siglip2-base-patch16-naflex?
At 4-bit, clip-vit-large-patch14 fits on 1x MI300X from $1.85 an hour and siglip2-base-patch16-naflex on 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.
Can I use siglip2-base-patch16-naflex commercially?
Yes. siglip2-base-patch16-naflex is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.