Points Rules
1. The general points given by subscribing to any plan apply to all models.
2. For models in the standard channel, points are consumed per use. The price list is as follows:
| Standard Channel Model | Price (points/use) |
|---|---|
| GPT-100(pro/pro/gpt-5.2-pro) | 112 |
| GPT-5.2 | 112 |
| GPT-5.1 (Response) | 100 |
| GPT-5.2-instant | 112 |
| GPT-5.4-thinking | 252 |
| GPT-5.2-thinking | 252 |
| Sora2 | 1000 |
| Seedance 2.0（全能） | 360 |
| Seedance 2.0（首尾帧） | 360 |
| Seedance 2.0（文生） | 360 |
| Seedance 2.0 Fast（全能） | 1200 |
| Seedance 2.0 Fast（首尾帧） | 1200 |
| Seedance 2.0 Fast（文生） | 1200 |
| Seeddream 4.0 | 120 |
| Seeddream 4.5 | 120 |
| Nano Banana Pro 2k | 200 |
| Nano Banana(偶有不稳) | 180 |
| Qwen3 Max | 180 |
| GPT-5（速度慢, 更智能） | 100 |
| Nano Banana2 | 60 |
| Nano Banana HD(10-20兆) | 180 |
| GPT Image 2 | 60 |
| GPT Image 2（异步） | 60 |
| GPT Image 2（Official） | 60 |
| GLM-4.5 air | 0 |
| Qwen Image | 100 |
| Nano Banana(人多不稳) | 80 |
| GPT-4o mini | 4 |
| GPT-5 Pro | 1600 |
| gpt-4o-image | 60 |
| GPT-4o | 6 |
| GPT-4.1 | 12 |
| Path of Exile | 60 |
| GPT-4.1 Mini | 2 |
| Claude Opus 4.5 thinking | 395 |
| Claude Opus 4.5 | 395 |
| Grok 4.1 | 80 |
| Nano Banana 2 2k | 200 |
| GPT-5.1 Pro | 1600 |
| GPT-5.2 Pro | 2240 |
| GPT-5.4 Pro | 1600 |
| Seedance 1.5 Pro (5s) | 200 |
| Seedance 1.5 Pro (12s) | 475 |
| Claude Opus 4.1 | 329 |
| Claude Opus 4 | 329 |
| o3 | 60 |
| Diablo | 80 |
| FLUX.1 Kontext pro | 173 |
| FLUX.1 Kontext max | 346 |
| DeepSeek V3 | 8 |
| DeepSeek R1 | 18 |
| DeepSeek V3 0324 | 8 |
| Qwen3 235B A22B | 5 |
| Qwen3 32B | 5 |
| QwQ 32B | 1 |
| GPT-4.1 Nano | 1 |
| o3 mini | 8 |
| o1 | 129 |
| Grok 3 | 27 |
| Grok 3 Reasoner | 40 |
| Gemini 3.1 Pro | 120 |
| Gemini 3 Pro Thinking | 120 |
| Gemini 2.0 Flash | 1 |
| Gemini 2.5 Pro | 36 |
| Claude Sonnet 4 | 5 |
| Claude 3.5 Sonnet | 5 |
| Claude 3.7 Sonnet | 36 |
| Qwen Max | 4 |
| Llama 4 Maverick | 1 |
| Llama 4 Scout | 1 |
3. For models in the high-speed stable channel, points are consumed based on the actual input and output tokens. The price list is as follows:
| High-Speed Stable Channel Model | Unit Price (points/ktokens) | |
|---|---|---|
| Input | Output | |
| GPT-4.1 | 9 | 36 |
| GPT-4o | 11 | 44 |
| GPT-image-1 | 66 | 176 |
| Grok 4 | 14 | 66 |
| o3 Pro | 88 | 351 |
| Claude Opus 4 | 66 | 329 |
| Gemini 2.5 Flash Preview | 1 | 3 |
| o3 | 44 | 176 |
| Claude Sonnet 4.0 | 14 | 66 |
| Claude 3.7 Sonnet | 14 | 66 |
| Gemini 2.5 Pro | 6 | 44 |
| Grok 3 | 14 | 66 |
| Grok 4 Fast | 2 | 5 |
- ① Token: A unit of text length unique to LLMs (large language models). The conversion between tokens and text length may vary by model. Generally, 1 token is equivalent to a word (English/French, etc.) or a character (Japanese/Chinese, etc.).
- ② Calculation formula:
Points consumed per use = Model input price × input tokens + Model output price × output tokens - ③ Calculation rule: If the tokens consumed are less than 1k, it will be calculated as 1k tokens.
Example:
In a conversation, you use the high-speed channel GPT-4o model, input 1600 tokens, and output 400 tokens. For this calculation, input is counted as 2k tokens, output as 1k tokens. The input price for GPT-4o is 11 points/ktokens, output price is 44 points/ktokens. Substitute into the formula:
Points consumed = 11 × 2 + 44 × 1 = 66 points
This conversation will consume 66 general points.