Model Pricing and Performance Guide
June 6, 2025
Model Credit Consumption Comparison
The following table shows the credit consumption of various AI models supported by FeelFish:
| Model Name | Credits per token (input/output) | Credits for ~100 words | Price Rating |
|---|---|---|---|
| DeepSeek-R1 | 55/219 | ≈200,000 | ⭐⭐⭐ (Moderate) |
| DeepSeek-V3 | 27/110 | ≈100,000 | ⭐⭐ (Economical) |
| Kimi-K2 | 48/192 | ≈180,000 | ⭐⭐ (Economical) |
| gpt-4.1 | 200/800 | ≈500,000 | ⭐⭐⭐⭐ (Expensive) |
| gpt-4o | 250/1000 | ≈550,000 | ⭐⭐⭐⭐ (Expensive) |
| qwen-plus | 11/28 | ≈25,000 | ⭐ (Very Economical) |
| gemini-2.5-pro | 125/1000 | ≈550,000 | ⭐⭐⭐⭐ (Expensive) |
| grok-3 | 300/1500 | ≈900,000 | ⭐⭐⭐⭐ (Expensive) |
Note: Credit consumption per word is estimated as context is included in the calculation.
- ⭐: Very Economical
- ⭐⭐: Economical
- ⭐⭐⭐: Moderate
- ⭐⭐⭐⭐: Expensive
- ⭐⭐⭐⭐⭐: Very Expensive
Model Use Cases and Writing Performance
Comparison of use cases and writing performance for each model:
| Model Name | Use Cases | Writing Performance |
|---|---|---|
| DeepSeek-V3 | Suitable for daily writing | Strong writing ability, fluent text |
| DeepSeek-R1 | Suitable for creative writing | Strong creativity |
| Kimi-K2 | Suitable for daily writing | Strong reasoning ability, fluent text |
| gpt-4.1 | Suitable for professional writing | - |
| gpt-4o | Suitable for professional creative writing | - |
| qwen-plus | Suitable for content summarization | Fast speed |
| gemini-2.5-pro | Suitable for multilingual writing | Strong multilingual and reasoning ability |
| grok-3 | Suitable for unique style scenarios | Unique and bold style |
Tip: Choose the appropriate model based on your writing needs and budget. For everyday use, DeepSeek-V3 is most suitable.