Configure Your Chatbot

AI Model

Choose the AI model your chatbot uses to generate replies. The model's multiplier sets how many credits it uses per 1,000 tokens.

Nivon provides 25+ models from OpenAI, Anthropic, Google, xAI, and other providers. Each model has a credit multiplier that determines how its usage is counted against your available credits.

Available models

Image

Choose from multiple models across OpenAI, Anthropic, Google, xAI, and other providers. Model choice affects the available model family and credit usage.

OpenAI

  • GPT-6.1 Sol · 12.25x
  • GPT-5.6 Terra · 6.25x
  • GPT-6 Luna · 2.5x
  • GPT-5.4 Mini · 2x
  • GPT-5.4 Nano · 0.75x
  • GPT-5 mini · 1x
  • GPT-5 Nano · 0.25x

Anthropic

  • Claude Opus 5.5 · 10.5x
  • Claude Sonnet 5.5 · 6.5x
  • Claude Haiku 5.5 · 2.25x

Google

  • Gemini 3.8 Flash · 3.25x
  • Gemini 3 Flash Preview · 1.25x
  • Gemini 3.5 Flash-Lite · 1x
  • Gemini 3.1 Flash-Lite · 0.75x
  • Gemma 4 · 0.25x

xAI

  • Grok 4.7 · 3x
  • Grok 4.3 · 1.5x

Other models

  • Kimi K3 (Moonshot) · 6.5x
  • GLM 5.3 (Zhipu) · 2.25x
  • DeepSeek V4.1 Flash · 1.75x
  • MiniMax M3 · 0.75x
  • Hy3 (Tencent Hunyuan) · 0.5x
  • Llama 4 Scout Instruct (Meta) · 0.5x
  • MiMo V2.5 (Xiaomi) · 0.25x

Understanding credit multipliers

Each model has a multiplier shown next to its name.

Credits are based on tokens: 1 credit = 1,000 tokens used on the base (1x) model. More advanced models use more credits per 1,000 tokens, according to their multiplier; lighter models with a multiplier below 1x use fewer.

For example:

  • 0.25x: 0.25 credits per 1,000 tokens.
  • 1x (base model): 1 credit per 1,000 tokens.
  • 2x: 2 credits per 1,000 tokens.
  • 6.25x: 6.25 credits per 1,000 tokens.

Your available credits and usage are shown in Credits and Usage. Each plan's monthly credit allowance is on the Pricing page.

Choosing a model

Choose a model based on the type of conversations your chatbot handles and the credit usage you want to maintain.

Consider:

  • Response requirements: More complex support conversations may require a different model than simple FAQ responses.
  • Expected conversation volume: Higher-volume chatbots can benefit from models with lower credit multipliers.
  • Testing results: Test the models with real customer questions before choosing one for production.

There is no single model that is the right choice for every chatbot. The best choice depends on your use case, expected usage, and test results.

Change the model

You can change the AI model from your chatbot configuration.

  1. Open your chatbot.
  2. Go to Chatbot Configuration.
  3. Open AI Model.
  4. Select a model.
  5. Save your changes.

Test the chatbot after changing the model to make sure its responses meet your requirements.

Last updated: October 11, 2026