Skip to content- Models are priced the same when using both the client and the API,However, the client will consume some additional tokens as system prompts。
- 使用 API 时模型名称与表格中模型名称相同但要全部小写,It doesn't matter which model,the data returned by the API is in OpenAI format。
Large language models
The name of the model | Enter a price (per 1K tokens/points) | Output price (per 1K tokens/points) | Input limit | Output limit |
---|
DeepSeek-Reasoner (R1) | 0.006 | 0.03 | 64k | 8k |
DeepSeek-Chat (V3) | 0.003 | 0.015 | 64k | 8k |
Grok-2 | 0.16 | 0.8 | 130k | 130k1 |
Grok-2-Vison | 0.16 | 0.8 | 32k | 32k1 |
Claude-3-7-Sonnet | 0.15 | 0.3 | 200k | 64k |
Claude-3-5-Sonnet | 0.15 | 0.3 | 200k | 4k |
Claude-3-Opus | 0.25 | 0.85 | 200k | 4k |
Claude-3-Haiku | 0.005 | 0.015 | 200k | 4k |
GPT-4.5-Preview | 0.8 | 1.7 | 128k | 16k |
GPT-4 | 0.40 | 0.70 | 8k | 8k |
GPT-4-Turbo | 0.15 | 0.35 | 128k | 4k |
GPT-4O | 0.03 | 0.15 | 128k | 4k |
GPT-4O-Mini | 0.002 | 0.009 | 128k | 4k |
GPT-3.5-Turbo | 0.005 | 0.015 | 16k | 4k |
Gemini-2.0-Flash | 0.005 | 0.01 | 1048k | 8k |
Gemini-2.0-Flash-Lite | 0.001 | 0.009 | 1048k | 8k |
Gemini-1.5-Pro | 0.05 | 0.20 | 2,000k | 8k |
Gemini-1.5-Flash | 0.003 | 0.009 | 1,000k | 8k |
O1 | 0.20 | 0.70 | 200k | 100k |
O1-Pro | 2 | 7 | 200k | 100k |
O3-Mini | 0.02 | 0.08 | 200k | 100k |
O1-Mini | 0.04 | 0.20 | 128k | 65k |
1: Grok 官方未给出具体参数。
Embedding model
The name of the model | price (per 1K tokens/points) | Token Limits |
---|
text-embedding-ada-002 | 0.001 | 8k |
text-embedding-3-large | 0.0013 | 8k |
text-embedding-3-small | 0.0002 | 8k |
Speech-related models
The name of the model | price (score) | limit |
---|
Whisper-1 | 0.06/second | 25MB |
GPT-4o-Transcribe | 0.06/second | 16KB |
GPT-4o-Mini-Transcribe | 0.03/second | 16KB |
TTS-1 | 0.15/1K characters | 4,096character |
TTS-1-HD | 0.30/1K characters | 4,096character |
GPT-4o-Mini-TTS | 0.15/1K characters | 2000character |