Service, Credit & Pricing

Service

GPTBots currently offers two different service modes, and customers can choose to use either the "GPTBots Key" or the "Own Key" according to their needs.

GPTBots Key: GPTBots officially provides the service directly, developers do not need to register for keys from platforms like OpenAI or Claude themselves, and can directly use their services on GPTBots;
Own Key: If developers have their own keys from platforms like OpenAI, they can also use them directly on the GPTBots platform, and GPTBots will charge a small amount of credits as a service fee.

You can choose and set the service mode you need within "Organization - LLMs".

Credit

All services within GPTBots are consumed using "credits".

Different services consume different amounts of credits, and the specific calculation of consumption is detailed below.

Note: Points are not refundable or exchangeable.

Pricing

LLM

Note: The following prices are measured in "credits / 1K Tokens".

Brand	Model	Iuput (Platform Key)	Output (GPTBots Key)	Iuput (Own Key)	Output (Own Key)
OpenAI	GPT-4o-mini-128k	0.0165	0.0665	0.0015	0.006
OpenAI	GPT-3.5-turbo-16k	0.055	0.165	0.005	0.015
OpenAI	GPT-4-8k	3.3	6.6	0.3	0.6
OpenAI	GPT-4-turbo-128k	1.1	3.3	0.1	0.3
OpenAI	GPT-4o-128k	0.55	1.65	0.05	0.15
Google	Gemini-1.5-Flash	0.039	0.116	0.03	0.15
Google	Gemini-1.5-Pro	0.385	1.106	0.035	0.106
Anthropic	Claude-3.0-Haiku-200k	0.028	0.138	0.003	0.013
Anthropic	Claude-3.0-Sonnet-200k	0.33	1.65	0.03	0.15
Anthropic	Claude-3.0-Opus-200k	1.65	8.25	0.15	0.75
Azure	GPT-3.5-turbo-16k	0.055	0.165	0.005	0.015
Azure	GPT-4-8k	3.3	6.6	0.3	0.6
Azure	GPT-4-32k	6.6	13.2	0.6	1.2
Meta	llama-3.0-8b-8k	0.028	0.028	0.002	0.002
Meta	llama-3.0-70b-8k	0.099	0.099	0.009	0.009
Mixtral	open-mistral-7b	0.028	0.028	0.003	0.003
Mixtral	open-mixtral-8x7b	0.077	0.077	0.007	0.007
Mixtral	mistral-small-latest	0.220	0.660	0.020	0.060
Mixtral	mistral-medium-latest	0.297	0.891	0.027	0.081
Mixtral	mistral-large-latest	0.880	2.640	0.080	0.240
Tencent	Hunyuan-lite- 4k	free	free	free	free
Tencent	Hunyuan-standard-32k	0.0707	0.0786	0.0064	0.0071
Tencent	Hunyuan-standard-256k	0.2357	0.9429	0.0214	0.0857
Tencent	Hunyuan-pro-32k	0.472	1.572	0.042	0.142
Ali	Qwen-turbo-8k	0.13	0.13	0.01	0.01
Ali	Qwen-plus-32k	0.32	0.32	0.02	0.02
Ali	Qwen-max-8k	1.88	1.88	0.17	0.17
Ali	Qwen-max-longcontext-30k	1.88	1.88	0.17	0.17
Ali	Qwen1.5-72b-32k	0.32	0.32	0.02	0.02
Ali	Qwen1.5-14b-8k	0.13	0.13	0.01	0.01
Ali	Qwen1.5-7b-8k	0.095	0.095	0.008	0.008
Baidu	ERNIE-4.0-8K	1.76	1.76	0.16	0.16
Baidu	ERNIE-3.5-8K	0.18	0.18	0.02	0.02
Baidu	ERNIE-Speed-128K	free	free	free	free
ZhiPu	GLM-3.0-6B-8K	0.095	0.095	0.008	0.008
ZhiPu	GLM-3.0-Turbo-128k	0.017	0.08	0.0015	0.0015
ZhiPu	GLM-4.0-128K	1.56	1.56	0.14	0.14

Embedding

Note: The following prices are measured in "credits / 1K Tokens".

Brand	Model	GPTBots Key	Own Key
OpenAI	text-embedding-ada-002	0.0120	0.0010
OpenAI	text-embedding-3-large	0.0156	0.0013
OpenAI	text-embedding-3-small	0.0024	0.0002

Rerank

Note: The following prices are measured in "credits / 1K Tokens".

Brand	Model	GPTBots Key	Own Key
Jina	reranker-v1-base-en	0.0022	0.0001
Jina	reranker-v1-turbo-en	0.0022	0.0001
Jina	reranker-v1-tiny-en	0.0022	0.0001
Baai	bce-rerank	0.0022	0.0001
NteEase	bgep-rerank	0.0022	0.0001

ASR

Note: The following prices are measured in "credits / 60 secs".

Brand	Model	GPTBots Key	Own Key
OpenAI	Whisper Large-V2	0.66	0.06
OpenAI	Whisper Large-V3	0.88	0.08

TTS

Note: The following prices are measured in "credits / 1000 chars".

Brand	Model	Platform Key	Own Key
OpenAI	TTS	1.65	0.15
Ali	Sambert	1.56	0.14

Vector Storage

Note: The following prices are measured in "credits / 1K Tokens/ day".

Service	Charge
Vector Storage	0.001

FAQ

Credits and Characters, how are they converted?

Taking the service of OpenAI's large language model (LLM) ChatGPT-3.5-turbo-16K (input) as an example:

$$ 1000 credits ≈ 2.5 million English words ≈ 1.25 million Chinese characters/Japanese kana/Korean Hangul $$

How are credits calculated when using image inputs?

When using the "GPT" model with image content, the calculation method for image Tokens is as follows:

Obtain the length and width values in "px", for example: "1024px * 1024px".
Calculate the "Tiles" value, which is obtained by dividing both the width and height by 512, rounding up, and then multiplying the two resulting values.
Calculate the "Tokens", using the formula "85 + 170 * Tiles".

The complete calculation formula is as follows:

$$ Tiles = ⌈(width÷512)⌉×⌈(height÷512)⌉ $$

$$ Tokens = 85 + 170 × Tiles $$

Python code is as follows:

import math def calculate_tokens(width, height): tiles = math.ceil(width/512) * math.ceil(height/512) tokens = 85 + 170 * tiles return tokens # Test print(calculate_tokens(2000, 500))

          import math

def calculate_tokens(width, height):
    tiles = math.ceil(width/512) * math.ceil(height/512)
    tokens = 85 + 170 * tiles
    return tokens

# Test
print(calculate_tokens(2000, 500))

このコードブロックは、フローティングウィンドウに表示されます

For example, if the input image dimensions are 2000px * 500px, then its Tiles value is 4*1=4, and the input Tokens for that image would be 85 + 170 * 4 = 765.