Find answers to your questions

What are AI credits and How it works?

Learn more about AI credits, Features, usage and more

1. Understanding AI Credits:

AI credits represent the usage of AI-powered features in your chatbot. An AI credit is calculated as the product of tokens and the AI model's consumption rate that is used when performing any AI task.

2. Explore AI-Powered Features:

Our platform offers a variety of AI-driven features that can enhance your chatbot’s capabilities, such as:

• Generating AI-based answers

• Automatic learning from past conversations

• Sentiment analysis and top intent analysis

• Summary generation and faster AI-assisted replies

• AI Helpdesk and more

3. Understand Model Pricing:

Different AI models have varying costs, For example, OpenAI pricing for the GPT-4.1 model is $8 USD per million tokens. Whereas in our platform, 1 million credits are roughly equivalent to $6.5 USD

As the cost of the GPT-4.1 model is higher than our baseline rate, we apply a pricing multiplier of 2.5x to align with provider pricing.

So total credits consumed = Tokens used × Model consumption rate.

As the provider cost goes down, we also reduce the pricing accordingly. For other models, you can check the difference here: AI Model & Consumption Rate

4. Calculating the AI Credits:

You can use our AI Credits Calculator to calculate your AI credit usage. This tool will give you an approximate number of AI-powered knowledgebase messages your chatbot can handle. There are no limits on the number of automated or static responses created in the chatbot studio, so those won’t impact your credits.

5. Get More Done with AI

With our AI-first platform, you can unlock real value for your business using AI—driving high efficiency in customer support, boosting sales, Internal workflow automation, and more.

AI Models & Usage

Available AI models, usage tiers, and credit consumption rates

CivicAi provides access to multiple AI models across different capability tiers. Each tier is designed to balance reasoning strength, speed, and credit usage so you can choose the right model for each task.


Tier 1 – Advanced Reasoning Models

Tier 1 models are the most capable and are best suited for high-accuracy tasks such as complex reasoning, planning, and critical decision flows.

Model

Consumption Rate

Function & Studio

Claude 4 Opus

15x

Yes

Claude 3.7 Sonnet

15x

Yes

DeepSeek R1

8x

No


Tier 2 – Balanced Performance Models

Tier 2 models offer a balance between intelligence, speed, and cost.
They are suitable for most production workloads and general AI agents.

Model

Consumption Rate

Function & Studio

GPT-5

2x

Yes

Kimi K2

1x

Yes

Grok-4

5x

Yes

GPT-4.1

2.5x

Yes

Claude Sonnet 4.5

3x

Yes

Claude Sonnet 4

3x

Yes

Grok-3

5x

Yes

DeepSeek V3

1x

No

Llama-4 Maverick

1x

Yes

Gemini 2.5 Pro Preview

4x

Yes

GPT-4o

2x

Yes

Meta Llama 3.1

3x

Yes

Mistral Large

4x

Yes


Tier 3 – Fast & Cost-Efficient Models

Tier 3 models are lightweight and optimized for speed and lower credit usage.
They work well for simple queries, automation, and high-volume interactions.

Model

Consumption Rate

Function & Studio

Claude Haiku 4.5

1x

Yes

Grok-3 Mini

1x

Yes

Gemini 2.5 Flash

0.5x

Yes

Gemini 2.0 Flash

0.5x

Yes

Llama-4 Scout

0.5x

Yes

GPT-5 Mini

0.5x

Yes

GPT-4.1 Mini

1x

Yes

GPT-4o Mini

0.5x

Yes


Consumption rates

Consumption rates represent relative credit usage.
Higher-tier models consume more credits due to increased reasoning depth and output quality.

Credit usage depends on multiple factors, including agent persona length, knowledgebase nodes, response size, translation, and function execution. For cost-sensitive workloads, start with Tier 2 or Tier 3 models and move to Tier 1 only when higher reasoning accuracy is required.

Was this article helpful?
©2026
Powered by CivicAI