Credit is a basic unit that applies exclusively to Vext Flow managed LLM usage. The total amount of allowed credits per account varies based on the plan you're on. You can learn more about the total credits you have for each plan here.
We decided not to use "token" because it could be confusing for some and difficult to manage/forecast usage and cost.
The consumption of credits per LLM query also depends on which LLM you choose, here is a chart that shows you how many credits per LLM query is consumed for each LLM:
Model | Credit | Plan |
---|---|---|
GPT 4o | 3 | Pro |
GPT 4o Mini | 1 | All |
GPT o1 Mini | 4 | Pro |
GPT o3 Mini | 1 | Pro |
Gemini Pro (1.5) | 2 | Pro |
Gemini 1.5 Flash | 1 | All |
Gemini 2.0 Flash | 1 | All |
Llama 3.2 3B | 1 | All |
Llama 3.2 90B | 1 | All |
Llama 3.3 70B | 1 | All |
Mistral 7B | 1 | All |
Mixtral 8x7B | 1 | All |
Mixtral 8x22B | 2 | Pro |
Mistral Small 3 | 1 | All |
Mistral Large | 2 | Pro |
Claude 3.5 Haiku | 1 | All |
Claude 3.5 Sonnet | 5 | Pro |
Claude 3 Opus | 11 | Pro |
Llama 3.1 Nemotron 70B | 1 | All |
DeepSeek R1 | 2 | Pro |
DeepSeek V3 | 1 | All |
Sonar | 3 | Pro |
Sonar Reasoning | 5 | Pro |
Qwen2 7B Instruct | 1 | All |
Qwen2.5 72B Instruct | 1 | All |
Qwen-Max | 4 | Pro |
Qwen-Plus | 2 | Pro |
Qwen-Turbo | 1 | All |
Phi 4 | 1 | All |
Grok 2 | 3 | Pro |
<aside> 💡
If you're bringing your own model to the platform, no credits will be used when the workflow is triggered.
</aside>