GPAI Ledger › Gemma 4 (Google) › Capture 11 Aug 2026

Gemma 4 — capture 20260811T103528Z

Verify: sha256sum 6e6966d9022d0a981d532d0eda75eabad26ecfd746214c35640bbbf82235ec59.pdf must equal the hash above (the filename IS the expected hash); ots verify 6e6966d9022d0a981d532d0eda75eabad26ecfd746214c35640bbbf82235ec59.pdf.ots -f 6e6966d9022d0a981d532d0eda75eabad26ecfd746214c35640bbbf82235ec59.pdf (opentimestamps.org) proves the capture time (fresh proofs report 'pending' until bitcoin-anchored, typically within a day).

Extracted text

Machine-extracted text (layout may be lost; the authoritative content is the stored file above).

Public  Summary
of

Training

Content

for

General-Purpose

AI

models

Version  of  the  Summary:  v1  Last  update:  31  July  2026

1.  General  information
1.1.  Provider  identification
Provider  name  and  contact  details:
Google  Ireland  Limited   Gordon  House  Barrow  Street  Dublin  4  Ireland
Authorised  representative  name  and  contact  details:

1.2.  Model  identification
Versioned  model  name(s):
The  Gemma  4  family  of  models

Model  dependencies: Gemma  4  is  not  a  modification  or  a  fine-tune  of  a  prior  model.  Each  subsequent  model  in  the  Gemma  4  family  is  based  on  Gemma  4  (see  each  model  card  for  individual  model  details).  The  Gemma  4  family  includes  models  such  as:  Gemma  4  E2B,  Gemma  4  E4B,  Gemma  4  12B,  Gemma  4  26B  A4B,  Gemma  4  31B.
Date  of  placement  of  the  model  on  the  Union  market:
April  2026
1.3  Modalities,  overall  training  data  size  and  other  characteristics
Modality  Training  data  size Types  of  content
 Text   Less  than  1  billion  tokens   1  billion  to  10  trillion  tokens   More  than  10  trillion  tokens
The  model  trains  on  datasets  containing  information  that  can  be  represented  as  written  language,  including  software  code  and  publicly  available  web  documents  across  educational,  government,  legal,  and  research  sectors.

2

 Image   Less  than  1  million  images   1  Million  to  1  billion  images   More  than  1  billion  images
The  model  trains  on  datasets  representing  static  visual  information,  including  photography,  diagrams,  and  illustrations.
 Audio

 Less  than  10  000  hours   10  000  to  1  million  hours    More  than  1  million  hours
The  model  trains  on  datasets  representing  sound  that  has  been  recorded  and  digitized,  including  speech  recordings  and  sound  effects.
 Video   Less  than  10  000  hours   10  000  to  1  million  hours    More  than  1  million  hours
The  model  trains  on  datasets  containing  a  sequence  of  image  frames  that  may  be  accompanied  by  audio,  including  performances,  video  clips,  and  video  effects.
 Other

Latest  date  of  data  acquisition/collection  for  model  training:
The  knowledge  cut-off  date  for  Gemma  4  is  January  2025.  We  may  periodically  update  models  after  this  date.
Description  of  the  linguistic  characteristics  of  the  overall  training  data:
Our  models  are  trained  to  work  with  a  breadth  of  languages,  including  EU  official  languages,  representative  of  the  internet  in  general.
Other  relevant  characteristics  of  the  overall  training  data:
The  pre-training  dataset  was  a  large-scale,  diverse  collection  of  data  encompassing  a  wide  range  of  domains  and  modalities,  which  included  publicly-available  web-documents,  code,  images,  audio  (including  speech  and  other  audio  types),  and  video.  The  post-training  dataset  included  different  types  of  instruction  tuning  data,  reinforcement  learning  data,  and  human-preference  data.
Additional  comments  (optional):

2.  List  of  data  sources
2.1.  Publicly  available  datasets
Have  you  used  publicly  available  datasets  to  train  the  model?
 Yes
  No

3

If  yes,  specify  the  modality(ies)  of  the  content  covered  by  the  datasets  concerned:
 Text   Image   Video
 Audio

Other:
List  of  large  publicly  available  datasets:
Our  publicly  available  datasets  include  data  across  various  sectors  such  as  educational,  government,  legal,  and  research  sectors  comprising  a  wide  variety  of  media  types  and  languages.

General  description  of  other  publicly  available  datasets  not  listed  above:
See  the  description  of  linguistic  characteristics  and  other  relevant  characteristics  of  overall  training  data  in  Section  1.
Additional  comments  (optional):

2.2  Private  non-publicly  available  datasets  obtained  from  third  parties
2.2.1.  Datasets  commercially  licensed  by  rightsholders  or  their  representatives
Have  you  concluded  transactional  commercial  licensing  agreement(s)  with  rightsholder(s)  or  with  their  representatives?
 Yes
  No

If  yes,  specify  the  modality(ies)  of  the  content  covered  by  the  datasets  concerned:
 Text   Image   Video
 Audio

 Other:

2.2.2.  Private  datasets  obtained  from other  third  parties
Have  you  obtained  private  datasets  from  third  parties  that  are  not  licensed  as  described  in  Section  2.2.1,  such  as  data  obtained  from  providers  of  private  databases,  or  data  intermediaries?
 Yes
  No

If  yes,  specify  the  modality(ies)  of  the  content  covered  by  the  datasets  concerned
 Text   Image   Video
 Audio

 Other:
If  publicly  known,  list  private  datasets  obtained  from  other  third  parties:
None  of  the  private  datasets  subject  to  section  2.2.2  are  publicly  known.

4

General  description  of  non-publicly  known  private  datasets  obtained  from  third  parties
Datasets  cover  subject  areas  including  educational  materials,  scientific  reasoning,  and  mathematics.
Additional  comments  (optional):

2.3  Data  crawled  and  scraped  from  online  sources
Were  crawlers  used  by  the  provider  or  on  behalf  of?   Yes
  No

If  yes,  specify  crawler  name(s)/identifier(s):
See  list  of  crawlers  here.
Purposes  of  the  crawler(s):
Google  uses  crawlers  to  discover  and  scan  websites,  find  information  for  building  Google's  search  indexes,  perform  other  product  specific  crawls,  and  for  analysis.
General  description  of  crawler  behaviour:
Google's  crawlers  are  used  to  to  discover  and  scan  websites,  find  information  for  building  Google's  search  indexes,  perform  other  product  specific  crawls,  and  for  analysis.  For  more  information  on  crawler  behaviour,  see  here.
Period  of  data  collection:
The  knowledge  cut-off  date  for  Gemma  4  is  January  2025.  We  may  periodically  update  models  after  this  date.
Comprehensive  description  of  the  type  of  content  and  online  sources  crawled:
Crawled  data  includes  a  broad  range  of  publicly  available  online  material,  such  as  publicly  available  websites  across  educational,  government,  legal,  and  research  sectors.  This  includes  a  wide  variety  of  media  types  and  languages.
Type  of  modality  covered:
 Text   Image   Video   Audio
 Other:
Summary  of  the  most  relevant  domain  names  crawled:
The  most  relevant  domains  crawled  include  publicly  available  websites  across  educational,  government,  legal,  and  research  sectors  comprising  a  wide  variety  of  media  types  and  languages.  Google  uses  crawlers  to  discover  and  scan  websites,  find  information  for  building  Google's  search  indexes,  perform  other  product  specific  crawls,  and  for  analysis.  See  list  of  our  common  crawlers  here.  Google’s  common  crawlers  obey  robots.txt  rules  when  crawling  automatically.   To  learn  more  about  Google’s  crawlers,  visit this  site.
Additional  comments  (optional):

2.4  User  data

5

Was  data  from  user  interactions  with  the  AI  model  (e.g.  user  input  and  prompts)  used  to  train  the  model?
 Yes
  No

Was  data  collected  from  user  interactions  with  the  provider’s  other  services  or  products  used  to  train  the  model?
 Yes
  No

If  yes,  provide  a  general  description  of  the  provider’s  services  or  products  that  were  used  to  collect  the  user  data:
In  accordance  with  Google’s  relevant  terms  of  service,  privacy  policy,  and  service-specific  policies,  and  pursuant  to  user  consent  where  legally  required,  we  use  data  collected  from  users  of  Google  products  and  services—such  as  user  chats  with  Gemini  Apps—to  train  AI  models.
Type  of  modality  covered:
 Text   Image   Video   Audio
 Other:
Additional  comments  (optional):

2.5  Synthetic  data
Was  synthetic  AI-generated  data  created  by  the  provider  or  on  their  behalf  to  train  the  model?
 Yes
  No

If  yes,  modality  of  the  synthetic  data:
 Text   Image   Video   Audio
 Other:
If  yes,  specify  the  general-purpose  AI  model(s)  used  to  generate  the  synthetic  data  if  available  on  the  market:
We  have  relied  on  the  externally  available Gemini  3  family of  models  to  generate  synthetic  data.
Information  about  other  AI  models,  including  provider’s  own  AI  model(s)  not  available  on  the  market,  used  to  generate  synthetic  data  to  train  the  model  to  which  this  Summary  applies:
We  may  use  fine-tuned  versions  of  internal  models  and  non-GPAI  models  to  generate  synthetic  data  for  training.

6

Additional  comments  (optional):

2.6  Other  sources  of  data
Have  data  sources  other  than  those  described  in  Sections  2.1  to  2.5  been  used  to  train  the  model?
 Yes   No
If  yes,  provide  a  narrative  description  of  these  data  sources  and  the  data:
This  category  reflects  types  of  data  that  do  not  fit  within  the  categories  described  above.  It  includes  datasets  that  Google  acquires  or  generates  in  the  course  of  its  business  operations,  or  directly  from  its  workforce.
Additional  comments  (optional):

3.  Data  processing  aspects
3.1.  Respect  of  reservation  of  rights  from  text  and  data  mining  exception  or  limitation
Are  you  a  Signatory  to  the  Code  of  Practice  for  general-purpose  AI  models  that  includes  commitments  to  respect  reservations  of  rights  from  the  TDM  exception  or  limitation?
 Yes   No
Describe  the  measures  implemented  before  model  training  to  respect  reservations  of  rights  from  the  TDM  exception  or  limitation  before  and  during  data  collection,  including  the  opt-out  protocols  and  solutions  honoured  by  the  provider  or,  as  applicable,  by  third  parties  from  which  datasets  have  been  obtained:
Data  filtering  and  preprocessing  included  techniques  such  as  deduplication,  honoring  robots.txt,  safety  filtering  in-line  with  Google's  commitment  to  advancing  AI  safely  and  responsibly,  and  quality  filtering  to  mitigate  risks  and  improve  training  data  reliability.  Specifically,  our  Google-Extended control  lets  web  publishers  manage  whether  content  Google  crawls  from  their  sites  may  be  used  for  training  Gemini  models  that  power  Gemini  Apps and  Gemini  Enterprise  Agent  Platform,  for  grounding  in  Gemini  Apps,  and  for  the  Grounding  with  Google  Search  feature  on  Gemini  Enterprise  Agent  Platform.  The  data  filtering  and  preprocessing  process  may  also  involve  filtering  irrelevant  or  harmful  content,  text,  and  other  modalities,  including  filtering  content  that  is  pornographic,  violent,  or  violative  of  child  sexual  abuse  material  (CSAM)  laws.

Additional  comments  (optional):

3.2  Removal  of  illegal  content

7

General  description  of  measures  taken:
As  described  above,  data  filtering  and  preprocessing  included  techniques  such  as  safety  filtering  in-line  with  Google's  commitment  to  advancing  AI  safely  and  responsibly.  This  process  may  involve  filtering  irrelevant  or  harmful  content,  text,  and  other  modalities,  including  filtering  content  that  is  pornographic,  violent,  or  violative  of  child  sexual  abuse  material  (CSAM)  laws.
3.3.  Other  information  (optional)
Other  relevant  information  about  data  processing  (optional):

8