GPT-5.5

GPAI Provider: OpenAI

Metadata

Model URL: https://openai.com/index/introducing-gpt-5-5/
When was this model published? 2025-08-07
Is this a new or a fine-tuned model? New model
When was the Public Summary last updated? 2026-06-26
When did we evaluate the Public Summary? 2026-07-09
Where did we find the Public Summary? We found this document linked on this page. An archived version can be found here.

Results of Evaluation

Section Transparency Usefulness General
Clarity Completeness Consistency Correctness Accessibility Comprehension Transparency Usefulness
Document 81.82 100.0 100.0 70.0 65.0 66.67 87.04 65.79
General information 47.5 68.06 100.0 94.44 N/A 90.91 67.39 90.91
Public Data Sources 0.0 81.98 33.33 100.0 0.0 15.79 50.22 7.69
Private Data Sources 0.0 27.27 100.0 100.0 N/A 100.0 34.69 100.0
Scraped/Crawled Data 94.57 59.49 100.0 65.59 0.0 34.48 73.11 24.1
User Data 53.12 72.22 100.0 100.0 100.0 0.0 74.14 50.0
Synthetic/Other Data 90.77 87.23 100.0 100.0 100.0 50.0 92.73 70.73
Data Processing 100.0 58.33 100.0 100.0 100.0 47.83 90.0 57.14
Sum 58.44 63.43 94.44 86.63 52.45 45.25 67.59 47.52
Grades C+D+

Evaluation Notes

If you update your public summary, please let us know so that we can update the evaluation and scores. We also welcome suggestions on improvements to this page, e.g. see FAQ section.

The public summary for GPT-5.5 can be found on OpenAI's site, linked on one of OpenAI's help pages. The summary matches the structure of the template, with points being lost primarily due to providing high-level information where more detailed information is encouraged. We assessed the summary's score to be 67.59% with Grade C+ for transparency, and 47.52% with Grade D+ for usefulness.

Suggested Improvements
The following are the suggested improvements based on using our our methodology where the public summary had issues related to the specified metric. The severity represents the extent of the issue, with low indicating aspects that could be fixable, and high representing missing information or requiring major changes.

  • Document
    • M1 Document does not indicate whether it is the latest version or outdated. Document does not provide link to other versions of the document. low
    • M4 Document does not provide a link to the authoritative source of the document. Date format slightly incorrect. medium
    • M5 Document is only linked in a single location; not very discoverable. Document is not provided in the same context as the model. medium
    • M6 Document does not clearly indicate changes from previous version. Document does not indicate where notice of updates or changes will be provided. medium
  • General information
    • M1 Contact details refer to general business address, generic contact point. Description of types of content describes goals for which content is included. Languages are not described exactly (i.e. as a list), EU languages are not fully mentioned. high
    • M2 No links to additional publicly available documentation are provided with versioned model name. National, regional, demographic specificities of training data not provided. medium
    • M4 Latest data of data acquisition/collection for model training is not provided exactly. low
    • M6 It is not indicated whether the model is continuously trained on new or dynamic data after the latest data of data acquisition/collection. low
  • Public Data Sources
    • M1 List of large publicly available datasets only describes CommonCrawl. General description of other publicly available information contains some unnecessary information. high
    • M2 List of large publicly available datasets only describes CommonCrawl. Descriptions do not contain info relating to approximate start and end dates of data collection. low
    • M3 List of large publicly available datasets only describes CommonCrawl. high
    • M5 List of large publicly available datasets only describes CommonCrawl. high
    • M6 Descriptions provided make use of language which is difficult to understand by someone without expertise. high
  • Private Data Sources
    • M1 General description of non-publicly known private datasets obtained from third parties contains some superfluous information. high
    • M2 General description of non-publicly known private datasets obtained from third parties lack recommended information. high
  • Scraped/Crawled Data
    • M1 Exact dates of data collection are not given. low
    • M2 Suggested information is not provided in the comprehensive description of the type of content and online sources crawled. medium
    • M4 Dates of data collection are not given in the appropriate format. medium
    • M5 Summary of the most relevant crawled domains not properly provided. high
    • M6 Comprehensive description of the type of content and online sources crawled contains some superfluous information. high
  • User Data
    • M1 General description of provider's services or products that were used to collect user data contains some links to external pages. medium
    • M2 General description of provider's services or products that were used to collect user data are found in various locations. medium
    • M6 General description of provider's services or products that were used to collect user data are found in various locations. high
  • Synthetic/Other Data
    • M1 Listed model used to general synthetic data is not necessarily uniquely identifiable low
    • M2 Models used to general synthetic data is not necessarily exhaustive. low
    • M6 Information about synthetic models does not include a general description of the training data of the models. medium
  • Data Processing
    • M2 General description of measures taken to remove illegal content describes measures as 'may includ[ing]' certain measures, not will including. medium
    • M6 Provider does not describe how they use datasets with opt-out mechanisms, nor how they ensure that they are up to date with user rights requests. high
  • Sum
    • M1 Document does not indicate whether it is the latest version or outdated. Document does not provide link to other versions of the document.Contact details refer to general business address, generic contact point. Description of types of content describes goals for which content is included. Languages are not described exactly (i.e. as a list), EU languages are not fully mentioned.List of large publicly available datasets only describes CommonCrawl. General description of other publicly available information contains some unnecessary information.General description of non-publicly known private datasets obtained from third parties contains some superfluous information.Exact dates of data collection are not given.General description of provider's services or products that were used to collect user data contains some links to external pages.Listed model used to general synthetic data is not necessarily uniquely identifiable medium
    • M2 No links to additional publicly available documentation are provided with versioned model name. National, regional, demographic specificities of training data not provided.List of large publicly available datasets only describes CommonCrawl. Descriptions do not contain info relating to approximate start and end dates of data collection.General description of non-publicly known private datasets obtained from third parties lack recommended information.Suggested information is not provided in the comprehensive description of the type of content and online sources crawled.General description of provider's services or products that were used to collect user data are found in various locations.Models used to general synthetic data is not necessarily exhaustive.General description of measures taken to remove illegal content describes measures as 'may includ[ing]' certain measures, not will including. medium
    • M3 List of large publicly available datasets only describes CommonCrawl. low
    • M4 Document does not provide a link to the authoritative source of the document. Date format slightly incorrect.Latest data of data acquisition/collection for model training is not provided exactly.Dates of data collection are not given in the appropriate format. low
    • M5 Document is only linked in a single location; not very discoverable. Document is not provided in the same context as the model.List of large publicly available datasets only describes CommonCrawl.Summary of the most relevant crawled domains not properly provided. medium
    • M6 Document does not clearly indicate changes from previous version. Document does not indicate where notice of updates or changes will be provided.It is not indicated whether the model is continuously trained on new or dynamic data after the latest data of data acquisition/collection.Descriptions provided make use of language which is difficult to understand by someone without expertise.Comprehensive description of the type of content and online sources crawled contains some superfluous information.General description of provider's services or products that were used to collect user data are found in various locations.Information about synthetic models does not include a general description of the training data of the models.Provider does not describe how they use datasets with opt-out mechanisms, nor how they ensure that they are up to date with user rights requests. high