Special offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy.

iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more

Google’s Gemini technical report describes a multimodal model family built through a cross-Google effort. Its contributor section names technical and program leads, groups contributors by role, and thanks people who helped prepare and review the report. The report also says that names within each role are not ordered by contribution.

What the Gemini report covers

The Gemini Team introduced its work with the description: “We present Gemini, a family of highly capable multimodal models developed at Google.” The report, “Gemini: A Family of Highly Capable Multimodal Models”, was posted on December 19, 2023. It presents Gemini 1.0 as a family trained jointly on image, audio, video, and text data.

Three models for different constraints

Model Intended role in the report
Gemini Ultra Highly complex tasks
Gemini Pro Performance and deployability at scale
Gemini Nano On-device applications

These descriptions are the report’s framing of the models, not a current product comparison.

Special offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy.

What the contributor section says

The report presents Gemini as a cross-Google effort involving Google DeepMind, Google Research, Bard/Assistant, Knowledge and Information, Core ML, Cloud, Labs, and other groups. Its contributor material separates people into role categories rather than treating the list as one undifferentiated author ranking.

Named leadership roles

  • Overall technical leads: Jeffrey Dean and Oriol Vinyals, identified as making equal contributions.
  • Overall post-training lead: Slav Petrov.
  • Program leads: Demis Hassabis and Koray Kavukcuoglu.
  • Gemini App program leads: Amar Subramanya and Sissie Hsiao.

Beyond those named roles, the report lists leads, core contributors, contributors, program leads, an overall post-training lead, and overall technical leads. It explicitly states that ordering within each role does not indicate contribution order, so the sequence of names should not be read as a ranking.

Acknowledgments

The report thanks named leads for preparing the report and acknowledges reviewers and colleagues for discussions and feedback. These statements document how the report describes contributions and thanks; they do not independently establish the precise work performed by each person beyond the report’s own role definitions.

How to read the report’s benchmark claims

The Gemini Team’s 2023 report says Gemini Ultra advanced the state of the art on 30 of the 32 benchmarks it examined. In its MMLU discussion, it reports 90.04% accuracy. Both figures are results reported by the paper for its evaluations, not a current independent comparison of Gemini products.

Special offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Support on Ko-Fi

How the report described access

The report distinguished chat-focused Gemini Apps models from developer-focused Gemini API models. In its 2023 description, it named Gemini and Gemini Advanced among app experiences, and Google AI Studio and Cloud Vertex AI as developer access paths. These are historical descriptions in the paper; the report does not establish present-day availability, product names, or access terms.

Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.