Model Credit Costs in Runcell
Complete guide to AI models available in runcell, including credit costs, capabilities, and context windows
Model Credit Costs
This document provides an overview of all AI models available in runcell, including their credit costs, capabilities, and technical specifications. Understanding model costs helps you choose the right model for your tasks while managing your credit usage effectively.
| Note: we are working on migrating to real token usage based system, in future, credits will be token based which is more accurate representing your real usage.
How Credit Costs Work
The credit cost for each request is calculated as:
Credit Cost = Base Credits × Credit Rate
- Base Credits: Configured per message (varies by feature)
- Credit Rate: Model-specific multiplier (shown in the table below)
Each model has a different credit rate multiplier, allowing you to balance cost with performance based on your needs. The credits rate we set is based on model provider's price per token
Credits Usage Table
| Model | Credit Rate |
|---|---|
| Anthropic | |
anthropic/claude-sonnet-4-5 | 2.5x |
anthropic/claude-haiku-4-5 | 0.8x |
anthropic/claude-opus-4-5 | 4.2x |
anthropic/claude-opus-4-0 | 12.5x |
anthropic/claude-opus-4-1 | 12.5x |
anthropic/claude-sonnet-4-0 | 2.5x |
anthropic/claude-3-5-haiku | 0.5x |
anthropic/claude-3-5-sonnet-20241022 | 2.5x |
anthropic/claude-3-7-sonnet-20250219 | 2.5x |
| OpenAI | |
openai/gpt-4o | 1.0x |
openai/gpt-4.1 | 0.8x |
openai/gpt-4o-mini | 0.2x |
| OpenRouter (OpenAI) | |
openrouter/openai/gpt-5.1 | 1.0x |
openrouter/openai/gpt-5.2 | 1.4x |
openrouter/openai/gpt-5 | 1.0x |
openrouter/openai/o3 | 1.0x |
gemini/gemini-2.5-pro | 1.6x |
google/gemini-3-pro-preview | 1.0x |
| Other Providers | |
moonshotai/kimi-k2-0905 | 0.5x |
deepseek/deepseek-v3.2-exp | 0.2x |
qwen/qwen3-coder-plus | 0.8x |
x-ai/grok-code-fast-1 | 0.2x |