Embeddings Customer Feedback - Quick Survey
Thank you for taking a few minutes to provide feedback on your embeddings usage. Your input is valuable to help us improve our models and services. This survey should take approximately 5 minutes to complete.
Sign in to Google to save your progress. Learn more
How do you plan to use the Embedding API? *
Primary Use Case *
Required
Use Case Examples (Optional): Please provide up to 3 brief examples that represent your primary use case (e.g., example query for retrieval, example text for classification).
Evaluation Datasets: Which datasets do you use or plan to use to evaluate your embedding models? *
Evaluation Metrics: Which metrics do you use or plan to use to evaluate your embedding models? *
Compared Models: What other embedding models have you experimented with for your use case?
Motivations for (Not) Using Our Models: Briefly, what factors influence your decision to use or not use our models?
Desired Improvements: In what specific areas would you like to see improvements in our models or services? *
Natural Languages Used: Which natural languages are you currently using with embeddings? Are some languages used more than others? Any plans to add more languages? *
Programming Languages for Code Embeddings
Python
Java
C++
JavaScript
C#
Ruby
PHP
Swift
Go
R
COBOL
SQL
Other
Yes
No
Clear selection
Do you utilize Flexible Dimensions (MRL) feature?
Clear selection
Embedding Dimensionality: What are the typical dimensions of the embeddings you work with? *
Performance Requirements - Throughput: What are your throughput requirements for embedding generation and usage (e.g., queries per second - QPS)? *
Performance Requirements - Tokens: What are your traffic requirements for embedding generation and usage (e.g., tokens per request)? *
Performance Requirements - Latency: What are your latency requirements for embedding generation and usage (in milliseconds - ms)? *

To understand your price sensitivity for our new Gemini embedding model, at 15c per million (input) tokens & no cost for output vectors would you:

a) Find it very attractive and likely increase usage? 
b) Consider it a fair price? 
c) Start to look for alternatives or reduce usage? 
d) Find it prohibitively expensive?
 *
Email (Optional): If you are open to follow-up questions, please provide your email.
Submit
Clear form
Never submit passwords through Google Forms.
This form was created inside Google.com. Privacy & Terms
Google Forms