GPAI Ledger › Inkling Small (Thinking Machines) › Capture 17 Aug 2026

Inkling Small — capture 20260817T084009Z

Verify: sha256sum 01df116716f944232d2275d2349f165160832917774d59e905c0079e1c158b20.pdf must equal the hash above (the filename IS the expected hash); ots verify 01df116716f944232d2275d2349f165160832917774d59e905c0079e1c158b20.pdf.ots -f 01df116716f944232d2275d2349f165160832917774d59e905c0079e1c158b20.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).

Summary  of  Training  Content  for
General-Purpose

AI

Model
  Version  of  the  Summary:   Version  1.0  Last  update:   July  30,  2026  General  information
1.  General  information   1.1.  Provider  identification  Provider  name  and  contact  details:
Thinking  Machines  Lab  Inc.  95  3rd  Street,  2nd  Floor,  San  Francisco,  CA  94103  Authorised  representative  name  and  contact  details:  EDPO  may  be  contacted  at  info@edpo.com
 1.2.  Model  identification   Versioned  model  name(s):  Inkling-Small  https://thinkingmachines.ai/model-card/inkling-small/)
Model  dependencies:
 N/A.  This  model  is  the  first  version  of  its  kind  released.  Date  of  placement  of  the  model  on  the  Union  market:  July  30,  2026
1.3  Modalities,  overall  training  data  size  and  other  characteristics
Modality
Training  data  size
Types  of  content
 ☒  Text
☐  Less  than  1  billion  tokens  ☐  1billion  to  10  trillions  tokens  ☒  More  than  10  trillions  tokens

Inkling-Small  was  trained  on  a  mixture  of  publicly  available  content,  content  acquired  through  partnerships,  synthetic  content,  and  generated  content,  including  general  web  content,  reference  materials,  technical  documentation,  source  code,  and  other  text,  curated  and  filtered.
☒  Image
☐  Less  than  1  million  images  ☐  1Million  to1  billion  images  ☒  More  than  1  billion  images
Inkling-Small  was  trained  on  a  mixture  of  publicly  available  images,  images  acquired  through  partnerships,  synthetic  images,  and  generated  images,  including  images  from  different  kinds  of  web  content,  curated  and  filtered.
☒  Audio
 1

☐  Less  than  10  000  hours  ☐  10  000  to1  million  hours   ☒  More  than  1  million  hours
Inkling-Small  was  trained  on  a  mixture  of  publicly  available  audio  and  audio  acquired  through  partnerships,  including  audio  from  the  public  web,  curated  and  filtered.
1
 Excluding  audio  that  is  part  of  video,  as  this  should  be  reported  under  the  “video”  modality  instead.  Furthermore,  the  Commission
understands

the

modality

of

‘audio’

to

include

‘speech’.

1
☒  Video
☐  Less  than  10  000  hours  ☐  10  000  to1  million  hours   ☒  More  than  1  million  hours
Inkling-Small  was  trained  on  a  mixture  of  publicly  available  video,  video  acquired  through  partnerships,  and  generated  video,  including  video  from  the  public  web,  curated  and  filtered.
☐  Other
Specify  the  modality  and  for  each  one  indicate  approximate  size  and  unit  of  measurement
N/A.
 Latest  date  of  data  acquisition/collection  for  model  training:
The  latest  date  of  data  collection  for  the  model  training  is  July  2026.  Thinking  Machines  Lab  collects  and  uses  datasets  to  train  and  improve  its  generative  AI  services  on  an  ongoing  basis.
Description  of  the  linguistic  characteristics  of  the  overall  training  data:
The  training  datasets  used  by  Thinking  Machines  Lab  are  multilingual,  with  languages  including  English.
Other  relevant  characteristics  of  the  overall  training  data:  N/A
Additional  comments  (optional):  N/A

2.  List  of  data  sources
2.  List  of  data  sou r c es   2.1.  Publicly  available  datasets     Have  you  used  publicly  available  datasets  to  train  the  model?    ☒ Yes      ☐  No If  yes,  specify  the  modality(ies)  of  the  content  covered  by  the  datasets  concerned:
☒ Text    ☒ Image    ☒ Video   ☒ Audio  ☐  Other
If

so,

please

specify…
 List  of  large  publicly  available  datasets:   Our  large  publicly  available  datasets  include  Common  Crawl  (https://commoncrawl.org/) General  description  of  other  publicly  available  datasets  not  listed  above:
 Thinking  Machines  Lab  used  various  publicly  available  image-captioning  datasets.  Thinking  Machines  Lab  also  used  various  publicly  available  datasets  containing  text,  images,  audio,  and/or  video  content.   Additional  comments  (optional):  N/A

2
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
If

so,

please

specify… 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  so,  please  specify… If  publicly  known,  list  private  datasets  obtained  from  other  third  parties:   N/A
General  description  of  non-publicly  known  private  datasets  obtained  from  third  parties
 Thinking  Machines  Lab  has  acquired  text,  image,  audio,  and  video  content  from  various  third  parties.  The  content  includes  items  in  the  public  domain  as  well  as  content  that  may  be  subject  to  intellectual  property  protection  in  some  jurisdictions.   Additional  comments  (optional):   N/A
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):  N/A
Purposes  of  the  crawler(s):  Crawlers  were  used  to  download  content  from  publicly  available  sources  from  the  internet  for  the  purpose  of  model  training.    General  description  of  crawler  behaviour:
 Thinking  Machines  Lab’s  policy  is  that  crawlers  should  not  circumvent  captchas,  password-protections,  or  other  access  controls,  and  respect  robots.txt.  Period  of  data  collection:  From  2025  to  2026
Comprehensive  description  of  the  type  of  content  and  online  sources  crawled:
 Crawlers  obtained  a  broad  variety  of  content  types,  including  text  and  images,  drawn  from  publicly  available  sources.  Content  types  include  publicly  available  webpages,  PDF  documents,  academic  papers,  and  code  repositories.  Content  includes  academic,  scientific,  mathematical,  code-related,  and  general-purpose  content  (in  English  and  other  languages).  Type  of  modality  covered:   ☒ Text    ☒ Image    ☐  Video   ☐  Audio
3
 ☐  Other
If

so,

please

specify…

Summary  of  the  most  relevant  domain  names  crawled:
The  most  relevant  domains  used  to  train  Inkling-Small  include  resources  and  repositories  spanning  academic,  scientific,  mathematical,  code-related,  and  general-purpose  (in  English  and  other  languages)  content.  Additional  comments  (optional):   N/A
2.4  User  data   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:
 N/A
Type  of  modality  covered:
☐  Text    ☐  Image    ☐  Video   ☐  Audio   ☐  Other
If

so,

please

specify…
 N/A
Additional  comments  (optional):  N/A
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

so,

please

specify…
  If  yes,  specify  the  general-purpose  AI  model(s)  used  to  generate  the  synthetic  data  if  available  on  the  market:  Thinking  Machines  Lab  models  and  third  party  models  were  used  for  various  tasks  in  the  synthetic  data  pipeline,  including,  for  example,  augmentation  and  evaluation,  and  to  generate  examples.
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:
Additional  comments  (optional):  N/A
  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
4
If  yes,  provide  a  narrative  description  of  these  data  sources  and  the  data:    We  work  with  vendors  and  experts  on  specialty  datasets,  e.g.  safety.  Additional  comments  (optional):  N/A
1.  Data  processing  aspects  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:
 Thinking  Machines  Lab’s  policy  is  that  any  crawlers  and  data  partners  should  engage  in  legal  collection  and  should  not  circumvent  captchas,  password-protections,  or  other  access  controls,  and  respect  robots.txt.  Additional  comments  (optional):  N/A
 3.2  Removal  of  illegal  content
General  description  of  measures  taken:
Thinking  Machines  Lab  takes  steps  to  avoid  the  collection  of  data  that  is,  or  is  likely,  illegal  data,  and  performs  some  filtering  to  remove  illegal  data.
3.3.  Other  information  (optional)  Other  relevant  information  about  data  processing  (optional):  N/A
5