Close



Frequently Asked Questions



Frequently Asked Questions

Clients



Clients

Users



Users

Data Sources



Data Sources

Data Chat



Data Chat

Insights



Insights

Google Search Console



Insight | Google Search Console

Google Business Profile (Formerly Google My Business)



Insight | Google Business Profile

Google Ads



Insight | Google Ads

Google Analytics 4



Insight | Google Analytics

Website



Insight | Website

Facebook Ads



Insight | Facebook Ads

Microsoft Ads



Insight | Microsoft Ads

Automation Jobs



Automation Jobs

Google Search Console



Automation | Google Search Console

Google Business Profile



Automation | Google Business Profile

Google Ads



Automation | Google Ads

Google Analytics 4



Automation | Google Analytics

Website



Automation | Website

Facebook Ads



Automation | Facebook Ads

Microsoft Ads



Automation | Microsoft Ads

KPI Forecasting



KPI Forecasting

Google Analytics 4



KPI Forecasting | Google Analytics

Google Ads



KPI Forecasting | Google Ads

Google Business Profile



KPI Forecasting | Google Business Profile

Facebook Ads



KPI Forecasting | Facebook Ads

Microsoft Ads



KPI Forecasting | Microsoft Ads

Google Sheets



KPI Forecasting | Google Sheets

Budget Tracking



Budget Tracking

Google Sheets



Budget Tracking | Google Sheets

Anomaly Detection



Anomaly Detection

Data Warehouse



Data Warehouse

Google Looker Studio



Data Warehouse | Google Looker Studio

Google BigQuery



Data Warehouse | BigQuery

Google Business Profile (Formerly Google My Business)



Data Warehouse | Google Business Profile

Google Search Console



Data Warehouse | Google Search Console

Facebook Ads



Data Warehouse | Facebook Ads

Social Posts

Social Posts | Google Business Profile

Local Listings



Local Listings

Google Business Profile (Formerly Google My Business)



Local Listings | Google Business Profile

Platforms



Local Listings | Platforms

Media



Local Listings | Media

Categories & Services



Local Listings | Categories & Services

Reviews



Reviews

Settings



Settings

Integrations



Integrations

MCP



Integrations | MCP

Google Looker Studio



Integrations | Google Looker Studio

Slack



Integrations | Slack

Zapier



Integrations | Zapier

Make



Integrations | Make

Webhook



Integrations | Webhook

Comments



Comments

Google Looker Studio



Comments | Slack

Recent Updates



Account Settings

Settings

Acknowledging and Resolving Insights

Insights

Add comment from Slack

Comments | Slack

Users

Data Chat

Data Chat

Anomalies Bulk Action

Anomaly Detection

Anomaly Algorithm

Anomaly Detection

Integrations | Webhook

Anomaly Detection Insight

Insights

Anomaly Detection Settings

Anomaly Detection

Anomaly Job Common Errors

Anomaly Detection

An overview of Local Listings

Local Listings

An overview of Reviews

Reviews

Attributes

Local Listings

Authentication

Integrations | Make

Automatic Schema Updates

Data Warehouse

Automation: Facebook Ads Destination URL Checker

Integrations | Webhook

Automation: Google Ads Change History

Integrations | Webhook

Automation: Google Ads Destination URL Checker

Integrations | Webhook

Automation: Google Analytics Change History

Integrations | Webhook

Automation: Google Business Profile Answer

Integrations | Webhook

Automation: Google Business Profile Change History

Integrations | Webhook

Automation: Google Business Profile Pending Updates

Integrations | Webhook

Automation: Google Business Profile Question

Integrations | Webhook

Automation: Google Review

Integrations | Webhook

Automation: Google Review Change History

Integrations | Webhook

Automation Job Common Errors

Automation Jobs

Automation: Microsoft Ads Destination URL Checker

Integrations | Webhook

Automation: Page Status Checker

Integrations | Webhook

Automation: Robots.txt Change Detection

Integrations | Webhook

Automations Bulk Action

Automation Jobs

Automation: Search Console Activity

Integrations | Webhook

Automation: Search Console Keyword Movement

Integrations | Webhook

Auto Resolving

Insights

Budget Campaign Dynamic Filter

Budget Tracking

Budget Tracking

Budget Tracking

Budget Tracking

Budget Tracking

Budget Tracking

Budget Tracker Webhook Payload

Integrations | Webhook

Bulk Action

Local Listings

Bulk Upload Images

Local Listings | Media

Calendar

Social Posts

Cancel Subscription

Settings

AI Models

‍
‍

Choose from multiple AI models to balance performance, speed, and AI Credit usage based on your analysis needs.

‍

What are AI models?

Also called large language models (LLMs), these models power Data Chat answers. Each one differs in reasoning depth, speed, multimodal support, and how many credits each million tokens of input and output consumes. Picking the right model is about matching the job—not always using the largest one.

‍

AI credits are how Data Chat measures how much you use each model. Providers charge by tokens—think of them as small chunks of text (your question, pasted context, the assistant's answer, and sometimes hidden "thinking" steps). The numbers below are AI credits per 1 million tokens, shown separately for what you send in and what the model sends back.

You don't need to count tokens yourself. In practice: longer prompts, big pasted reports, and longer answers use more credits. Pick a cheaper band for quick tasks; use higher bands when you need deeper reasoning or polished long-form output (e.g. campaign narratives, complex cross-channel reads).

Why this matters for marketing teams: the tables help you compare models the same way you compare CPMs or software tiers—rough cost for a "unit" of usage—so you can default to efficient models for everyday questions and step up only when the brief is strategic or ambiguous.

How to read these numbers

These columns appear in the pricing tables. "Per 1M" means per one million tokens of that type. Your plan includes a monthly quota; additional credits can be purchased at $10 per 1,000 credits.

Column	What it means for you
Input (per 1M)	Credits charged for the text the model reads: your message, conversation history, and anything you paste in (briefs, performance notes, etc.). Larger prompts or long threads cost more on this line.
Output (per 1M)	Credits for the model's reply. Longer answers cost more. Some models bill extra internal "thinking" tokens here—you pay for quality and depth, not just the final paragraph.
Band	A simple label in the product (very low → very high) so you can compare models at a glance before you dive into the numbers.

Anthropic

Model	Input / 1M	Output / 1M	Band
Claude Haiku 4.5	110	550	Medium
Claude Sonnet 4.6	330	1650	Very high
Claude Opus 4.6	550	2750	Very high

Google

Model	Input / 1M	Output / 1M	Band
Gemini 3.1 Flash Lite Preview	28	165	Low
Gemini 3 Flash Preview	55	330	Low
Gemini 3.1 Pro Preview	220	1320	High

For some Google models, extra-long prompts can move you to a higher price tier than the standard row shown here. The figures match what Data Chat uses for typical prompt lengths.

OpenAI

Model	Input / 1M	Output / 1M	Band
GPT-5.4 Nano	22	138	Low
GPT-5.4 Mini	83	495	Medium
GPT-5.4	275	1650	Very high

Amazon

Model	Input / 1M	Output / 1M	Band
Nova 2 Lite	35	289	Low

Model	Input / 1M	Output / 1M	Band
Llama 4 Scout	12	37	Very low

xAI

Model	Input / 1M	Output / 1M	Band
Grok 4 Fast	22	55	Very low
Grok 4.20 Beta	220	660	Medium

Credit bands — what to use when

Think of bands as a quick shorthand. Match the job you're doing in Data Chat to the row that sounds like you—then pick any model in that band from the tables above.

Band	When marketers usually choose this
Very low	Quick metric checks, one-line clarifications, "what does this mean?" questions, or when you're testing prompts many times in a row.
Low	Standard campaign or channel questions, short summaries of performance, first drafts of bullets for stakeholders, everyday SEO/ads/email explanations.
Medium	Deeper but still focused work: comparing channels with context, outlining a report structure, interpreting a few metrics at once, or answers that need a bit more nuance.
High	Harder trade-offs (budget vs. reach, narrative for leadership), multi-part reasoning, or when you want stronger synthesis across messy or incomplete information.
Very high	The heaviest asks: long-form strategy framing, rigorous narrative for execs, complex multi-step analysis, or when you're okay paying more for the strongest model output.

‍

Model names, rates, and availability can change as providers and Jepto update the catalog. Preview models may change behavior or pricing alignment.

Tool use, grounding, and extended thinking can add tokens beyond a plain Q&A, which increases credits for the same user question.

Provider-specific rules (e.g. long-context surcharges) may apply; the usage chart is the source of truth for your consumption in Data Chat.

‍

Was this article helpful?



Nice one!

Thanks a lot for your feedback! If you’d like a member of our support team to respond to you, please send a message here



Please try again

Oops! Something went wrong while submitting the form.



Nice one!

Thanks a lot for your feedback! If you’d like a member of our support team to respond to you, please send a message here



Please try again

Oops! Something went wrong while submitting the form.

Help Center

AI Models

What are AI models?

How to read these numbers

Anthropic

Google

OpenAI

Amazon

Meta

xAI

Credit bands — what to use when