DEVELOPER REFERENCE · VERIFIED MODEL EVIDENCE

Compare 35 source-verified model records, repository licenses, published weight formats and dated evidence. No overall model ranking.

Evidence statusSource-first verified
Latest record check2026-11-13
Latest record full verification2026-11-04

DIRECT ANSWER · SOURCE-OBSERVED REGISTRY

What Google model evidence is available?

Open Model Weights currently tracks 35 field-verified Google model records. In this indexed sample, the recorded modalities are 8 Image + text → text; 7 Zero Shot Image Classification; 5 Any To Any; 5 Not declared; 3 Text; 2 Time Series Forecasting; 1 Image; 1 Image Feature Extraction; 1 Text + image → text; 1 Text embeddings; 1 Zero Shot Object Detection. The published repository license labels include 25 apache-2.0; 7 gemma; 3 other. The listed weight formats include Safetensors (33), PyTorch (9).

Coverage describes the model repositories indexed here, not every model released by the publisher. A checked source, an unknown field and an unchanged source revision are different states. No missing value is treated as a negative answer.

35verified model records
11recorded modalities
3distinct license labels
35model repository links

DEVELOPER HQ / PRIMARY ORGANIZATION LOCATION

Primary organization location

Mountain View, California, United States

Google

This location describes the developer or primary organization. It is not the nationality of a model, dataset, contributor, training run or infrastructure.

SOURCE-BACKED PUBLISHER PROFILE

Google: models, research and access

Google DeepMind publishes the Gemma open-weight family and separate open artifacts for embeddings, computer vision and forecasting. Google's official Gemma 4 documentation describes dense, Mixture-of-Experts and edge-oriented architectures with specific weight licenses.

Google's open-weight Gemma family is not the same as Google's proprietary Gemini API products. This registry also tracks research models such as SigLIP, ViT and TimesFM, which have different tasks, architectures and sometimes distinct licensing requirements.

01

Gemma 3 compared with Gemma 4

Google documents Gemma 3 and Gemma 4 as separate model generations. Gemma 4 includes dense 31B, MoE 26B A4B, 12B and smaller E2B/E4B multimodal options. The exact model card identifies the input types and recommended processors. Do not transfer Gemma 4 licensing or capability claims to Gemma 3 automatically.

02

License changes matter across generations

Google's published Gemma 3 Terms of Use differ from the Apache-2.0 license displayed on the Gemma 4 model card. A registry field saying gemma and one saying apache-2.0 cannot be treated as identical rights. Check model-specific usage, derivative, redistribution and responsible-use terms before deployment.

03

Beyond chat: embeddings, vision and time series

EmbeddingGemma records target semantic embedding and retrieval. SigLIP and ViT serve image recognition or feature extraction. TimesFM is a time-series forecasting family, with task-specific input handling. These weights must be assessed with different accuracy metrics and runtime software from general-purpose Gemma language models.

04

Open-weight Gemma versus hosted Gemini

A downloadable Gemma model can be deployed on infrastructure the operator chooses, subject to its terms. Gemini API access is a distinct service and should not be listed as a Gemma weight artifact. Claims about residency, costs or hosted data processing require evidence about the actual provider, region and deployment.

Publisher background is attributed to official sources. Model counts and license classifications come from OMW repository evidence; no model benchmark or processing location is inferred.

MODEL-BY-MODEL REFERENCE

Which Google open-weight model family should you inspect?

Browse by published modality and licensing evidence. These are navigation cues, not quality, speed or suitability rankings.

Text + image → textgemma

Gemma 3 27B IT

Gemma 3 multimodal checkpoint

Review the Gemma 3 terms and the model-specific image-plus-text support.

Commercial use: unknownLicense evidence →
Image + text → textapache-2.0

gemma-4-31B-it

Dense Gemma 4 model

Check the Gemma 4 31B variant, Apache-2.0 evidence and available compute.

Commercial use: allowedLicense evidence →
Image + text → textapache-2.0

gemma-4-26B-A4B-it

MoE Gemma 4 model

Compare active versus total parameters and serving requirements without assuming benchmark superiority.

Commercial use: allowedLicense evidence →
Text embeddingsgemma

embeddinggemma-300m

On-device or retrieval embeddings

Inspect EmbeddingGemma as a vector encoder with Gemma-specific license evidence.

Commercial use: unknownLicense evidence →

RELATED RESEARCH & RELEASES

Source-backed publisher coverage

News about additional products is not automatically a field-verified registry addition.

FIELD-BY-FIELD COMPARISON · PRIMARY LINKS

Model evidence at a glance.

Compare the exact entries in our registry. Parameter values are taken from the source-checked record; a license label is not a legal opinion, and recognized weight files are not an inference-speed benchmark. Open each record for the underlying citations and observed history.

Model and parameter countModalityLicense evidenceFormats and filesSource and history
vit-base-patch16-2240.0865677B published parametersImageapache-2.0allowed commercial-use classificationInspect license evidence →PyTorch, Safetensors2 filesRepository ↗Model evidence →History →
gemma-4-31B-it31.2731B published parametersImage + text → textapache-2.0allowed commercial-use classificationInspect license evidence →Safetensors3 filesRepository ↗Model evidence →History →
gemma-4-26B-A4B-it25.8059B published parametersImage + text → textapache-2.0allowed commercial-use classificationInspect license evidence →Safetensors3 filesRepository ↗Model evidence →History →
gemma-4-E4B-it7.99616B published parametersAny To Anyapache-2.0allowed commercial-use classificationInspect license evidence →Safetensors1 filesRepository ↗Model evidence →History →
gemma-4-E2B-it5.12318B published parametersAny To Anyapache-2.0allowed commercial-use classificationInspect license evidence →Safetensors1 filesRepository ↗Model evidence →History →
siglip2-giant-opt-patch16-3841.87189B published parametersZero Shot Image Classificationapache-2.0allowed commercial-use classificationInspect license evidence →Safetensors3 filesRepository ↗Model evidence →History →
timesfm-2.5-200m-pytorch0.231289B published parametersTime Series Forecastingapache-2.0allowed commercial-use classificationInspect license evidence →Safetensors1 filesRepository ↗Model evidence →History →
gemma-4-12B-it11.9597B published parametersAny To Anyapache-2.0allowed commercial-use classificationInspect license evidence →Safetensors1 filesRepository ↗Model evidence →History →
owlv2-base-patch16-ensemble0.154967B published parametersZero Shot Object Detectionapache-2.0allowed commercial-use classificationInspect license evidence →PyTorch, Safetensors2 filesRepository ↗Model evidence →History →
siglip2-base-patch16-2240.375188B published parametersZero Shot Image Classificationapache-2.0allowed commercial-use classificationInspect license evidence →Safetensors1 filesRepository ↗Model evidence →History →
siglip2-base-patch16-naflex0.375234B published parametersZero Shot Image Classificationapache-2.0allowed commercial-use classificationInspect license evidence →Safetensors1 filesRepository ↗Model evidence →History →
embeddinggemma-300m0.302863B published parametersText embeddingsgemmaunknown commercial-use classificationInspect license evidence →Safetensors3 filesRepository ↗Model evidence →History →
gemma-3-4b-it4.30008B published parametersImage + text → textgemmaunknown commercial-use classificationInspect license evidence →Safetensors3 filesRepository ↗Model evidence →History →
siglip-base-patch16-2240.203156B published parametersZero Shot Image Classificationapache-2.0allowed commercial-use classificationInspect license evidence →PyTorch, Safetensors2 filesRepository ↗Model evidence →History →
vit-base-patch16-224-in21k0.0863892B published parametersImage Feature Extractionapache-2.0allowed commercial-use classificationInspect license evidence →PyTorch, Safetensors2 filesRepository ↗Model evidence →History →
siglip-so400m-patch14-3840.87796B published parametersZero Shot Image Classificationapache-2.0allowed commercial-use classificationInspect license evidence →Safetensors1 filesRepository ↗Model evidence →History →
siglip2-so400m-patch14-3841.13601B published parametersZero Shot Image Classificationapache-2.0allowed commercial-use classificationInspect license evidence →Safetensors1 filesRepository ↗Model evidence →History →
flan-t5-base0.247578B published parametersNot declaredapache-2.0allowed commercial-use classificationInspect license evidence →PyTorch, Safetensors2 filesRepository ↗Model evidence →History →
mobilebert-uncasedPending verification published parametersNot declaredapache-2.0allowed commercial-use classificationInspect license evidence →PyTorch1 filesRepository ↗Model evidence →History →
diffusiongemma-26B-A4B-it25.8238B published parametersImage + text → textapache-2.0allowed commercial-use classificationInspect license evidence →Safetensors12 filesRepository ↗Model evidence →History →
gemma-4-12B-it-qat-w4a16-ct13.307B published parametersAny To Anyapache-2.0allowed commercial-use classificationInspect license evidence →Safetensors1 filesRepository ↗Model evidence →History →
timesfm-3.0-pytorch0.330711B published parametersTime Series Forecastingotherunknown commercial-use classificationLicense file ↗Safetensors1 filesRepository ↗Model evidence →History →
siglip2-base-patch16-2560.375234B published parametersZero Shot Image Classificationapache-2.0allowed commercial-use classificationInspect license evidence →Safetensors1 filesRepository ↗Model evidence →History →
gemma-4-E4B7.99616B published parametersAny To Anyapache-2.0allowed commercial-use classificationInspect license evidence →Safetensors1 filesRepository ↗Model evidence →History →
medgemma-4b-it4.30008B published parametersImage + text → textotherunknown commercial-use classificationInspect license evidence →Safetensors3 filesRepository ↗Model evidence →History →
gemma-3-12b-it12.1873B published parametersImage + text → textgemmaunknown commercial-use classificationInspect license evidence →Safetensors6 filesRepository ↗Model evidence →History →
Gemma 3 27B IT27B published parametersText + image → textgemmaunknown commercial-use classificationInspect license evidence →Safetensors13 filesRepository ↗Model evidence →History →
flan-t5-large0.78315B published parametersNot declaredapache-2.0allowed commercial-use classificationInspect license evidence →PyTorch, Safetensors2 filesRepository ↗Model evidence →History →
medgemma-1.5-4b-it4.30008B published parametersImage + text → textotherunknown commercial-use classificationInspect license evidence →Safetensors3 filesRepository ↗Model evidence →History →
flan-t5-small0.0769612B published parametersNot declaredapache-2.0allowed commercial-use classificationInspect license evidence →PyTorch, Safetensors2 filesRepository ↗Model evidence →History →
gemma-3-1b-it0.999886B published parametersTextgemmaunknown commercial-use classificationInspect license evidence →Safetensors1 filesRepository ↗Model evidence →History →
gemma-2-2b-it2.61434B published parametersTextgemmaunknown commercial-use classificationInspect license evidence →Safetensors3 filesRepository ↗Model evidence →History →
gemma-2-9b-it9.24171B published parametersTextgemmaunknown commercial-use classificationInspect license evidence →Safetensors5 filesRepository ↗Model evidence →History →
electra-base-discriminatorPending verification published parametersNot declaredapache-2.0allowed commercial-use classificationInspect license evidence →PyTorch1 filesRepository ↗Model evidence →History →
gemma-4-31B32.6824B published parametersImage + text → textapache-2.0allowed commercial-use classificationInspect license evidence →Safetensors3 filesRepository ↗Model evidence →History →

LICENSE & COVERAGE · NOT LEGAL ADVICE

How do the recorded licenses differ?

Different models from the same developer can use different license terms. The distribution below describes the labels and field-level classifications observed in the listed repositories; it does not automatically establish permissions for a specific deployment.

License labels

  • apache-2.025 of 35
  • gemma7 of 35
  • other3 of 35

Commercial-use classifications

  • allowed25 of 35
  • unknown10 of 35

Model modalities

  • Image + text → text8 of 35
  • Zero Shot Image Classification7 of 35
  • Any To Any5 of 35
  • Not declared5 of 35
  • Text3 of 35
  • Time Series Forecasting2 of 35
  • Image1 of 35
  • Image Feature Extraction1 of 35
  • Text + image → text1 of 35
  • Text embeddings1 of 35
  • Zero Shot Object Detection1 of 35

A classification of unknown is not a prohibition. Even a recognized permissive license can carry conditions. Review the license for the exact model and revision. Open the license explorer →

EVIDENCE METHOD · TRANSPARENT LIMITS

How this developer profile is assembled.

We aggregate the current source-checked model records, rather than publish an unverified corporate biography. Repository metadata, model cards, license files and recognized weight artifacts supply the facts where available. The record-check date does not assert that a model or its license changed on that date.

Source-derived runtime signals are not hardware benchmarks. Weight-only memory estimates omit runtime overhead, activations and other deployment requirements. Model-specific uncertainties remain visible in the individual evidence records.

QUESTIONS · DIRECT ANSWERS

Frequently asked questions about Google models.

Which Google models are covered?

Open Model Weights currently lists 35 field-verified records for Google. Their recorded model modalities are: 8 Image + text → text; 7 Zero Shot Image Classification; 5 Any To Any; 5 Not declared; 3 Text; 2 Time Series Forecasting; 1 Image; 1 Image Feature Extraction; 1 Text + image → text; 1 Text embeddings; 1 Zero Shot Object Detection. This is a registry sample, not an exhaustive publisher catalog.

Which licenses are recorded for Google models?

In the 35 indexed records, repository license labels are: 25 apache-2.0; 7 gemma; 3 other. License labels do not replace the model-specific terms or establish rights for every use case.

Where can I verify Google model information?

Each indexed model links to its official source repository and a field-level Open Model Weights profile. The most recent record checks shown here reach 2026-11-13; checking a repository does not imply that its contents changed.

Is Gemma 4 under the same license as Gemma 3?

No. Google's Gemma 4 model card identifies Apache 2.0, while its Gemma 3 models are governed by Gemma Terms of Use. Verify the exact model and attached terms.

Is Google's Gemini API model a downloadable Gemma checkpoint?

No. Google's Gemma documentation describes downloadable open weights. API-only Gemini products have different distribution and service arrangements and are not a Gemma repository artifact.

Can I compare TimesFM with Gemma 4 using chat benchmarks?

Not meaningfully. TimesFM is a forecasting model and Gemma 4 targets generative language and multimodal tasks. Use the relevant task and official model card.

COMPLETE INDEXED SELECTION

All 35 indexed model profiles.

vit-base-patch16-224
Image
Not declared in checked config/model card
apache-2.0
Field-verified
gemma-4-31B-it
Image + text → text
262,144 tokens
apache-2.0
Field-verified
gemma-4-26B-A4B-it
Image + text → text
262,144 tokens
apache-2.0
Field-verified
gemma-4-E4B-it
Any To Any
131,072 tokens
apache-2.0
Field-verified
gemma-4-E2B-it
Any To Any
131,072 tokens
apache-2.0
Field-verified
siglip2-giant-opt-patch16-384
Zero Shot Image Classification
Not declared in checked config/model card
apache-2.0
Field-verified
timesfm-2.5-200m-pytorch
Time Series Forecasting
16,384 tokens
apache-2.0
Field-verified
gemma-4-12B-it
Any To Any
262,144 tokens
apache-2.0
Field-verified
owlv2-base-patch16-ensemble
Zero Shot Object Detection
Not declared in checked config/model card
apache-2.0
Field-verified
siglip2-base-patch16-224
Zero Shot Image Classification
Not declared in checked config/model card
apache-2.0
Field-verified
siglip2-base-patch16-naflex
Zero Shot Image Classification
Not declared in checked config/model card
apache-2.0
Field-verified
embeddinggemma-300m
Text embeddings
2,048 tokens
gemma
Field-verified
gemma-3-4b-it
Image + text → text
128,000 tokens
gemma
Field-verified
siglip-base-patch16-224
Zero Shot Image Classification
Not declared in checked config/model card
apache-2.0
Field-verified
vit-base-patch16-224-in21k
Image Feature Extraction
Not declared in checked config/model card
apache-2.0
Field-verified
siglip-so400m-patch14-384
Zero Shot Image Classification
Not declared in checked config/model card
apache-2.0
Field-verified
siglip2-so400m-patch14-384
Zero Shot Image Classification
Not declared in checked config/model card
apache-2.0
Field-verified
flan-t5-base
Not declared
512 tokens
apache-2.0
Field-verified
mobilebert-uncased
Not declared
512 tokens
apache-2.0
Field-verified
diffusiongemma-26B-A4B-it
Image + text → text
262,144 tokens
apache-2.0
Field-verified
gemma-4-12B-it-qat-w4a16-ct
Any To Any
262,144 tokens
apache-2.0
Field-verified
timesfm-3.0-pytorch
Time Series Forecasting
Not declared in checked config/model card
other
Field-verified
siglip2-base-patch16-256
Zero Shot Image Classification
Not declared in checked config/model card
apache-2.0
Field-verified
gemma-4-E4B
Any To Any
131,072 tokens
apache-2.0
Field-verified
medgemma-4b-it
Image + text → text
128,000 tokens
other
Field-verified
gemma-3-12b-it
Image + text → text
128,000 tokens
gemma
Field-verified
Gemma 3 27B IT
Text + image → text
128,000 tokens
gemma
Field-verified
flan-t5-large
Not declared
512 tokens
apache-2.0
Field-verified
medgemma-1.5-4b-it
Image + text → text
128,000 tokens
other
Field-verified
flan-t5-small
Not declared
512 tokens
apache-2.0
Field-verified
gemma-3-1b-it
Text
128,000 tokens
gemma
Field-verified
gemma-2-2b-it
Text
Not declared in checked config/model card
gemma
Field-verified
gemma-2-9b-it
Text
Not declared in checked config/model card
gemma
Field-verified
electra-base-discriminator
Not declared
512 tokens
apache-2.0
Field-verified
gemma-4-31B
Image + text → text
262,144 tokens
apache-2.0
Field-verified