4,154 个模型 · 8,429 个价格 · 最近采集 6 小时前 · 30 天内 36 条需关注的调价
每个平台公布的每一个价格
采集器定时读取 6 个平台公开的全部价格（默认每天一次），并保存读到的原始页面。没有任何手工录入。搜索一个模型，就能看到现在按单位谁最便宜。
模型
Gemini 
Claude 
Qwen 
DeepSeek 
GLM 
Mistral 
Kimi 
Llama 
MiniMax 
Hunyuan 
Nemotron 第 1–51 条，共 51 条
每百万输入 token24
| 平台 | 模型 · 计价项 | 美元 | 平台原文 | 生效于 |
|---|---|---|---|---|
| OpenRouter | Meta: Llama 3.2 1B Instruct | $0.027/百万输入最低 | prompt: 0.000000027 / token | |
| OpenRouter | Sao10K: Llama 3 8B Lunaris | $0.04/百万输入 | prompt: 0.00000004 / token | |
| OpenRouter | Meta: Llama 3.1 8B Instruct | $0.05/百万输入 | prompt: 0.00000005 / token | |
| OpenRouter | Meta: Llama 3.2 3B Instruct | $0.05/百万输入 | prompt: 0.00000005 / token | |
| Replicate | meta/llama-2-7b | $0.05/百万输入 | $0.05 per million input tokens | |
| Replicate | meta/llama-2-7b-chat | $0.05/百万输入 | $0.05 per million input tokens | |
| Replicate | meta/meta-llama-3-8b | $0.05/百万输入 | $0.05 per million input tokens | |
| Replicate | meta/meta-llama-3-8b-instruct | $0.05/百万输入 | $0.05 per million input tokens | |
| OpenRouter | Meta: Llama 4 Scout | $0.1/百万输入 | prompt: 0.0000001 / token | |
| Replicate | meta/llama-2-13b | $0.1/百万输入 | $0.10 per million input tokens | |
| Replicate | meta/llama-2-13b-chat | $0.1/百万输入 | $0.10 per million input tokens | |
| Replicate | meta/llama-4-scout-instruct | $0.17/百万输入 | $0.17 per million input tokens | |
| OpenRouter | Meta: Llama Guard 4 12B | $0.18/百万输入 | prompt: 0.00000018 / token | |
| Replicate | meta/llama-guard-4-12b | $0.2/百万输入 | $0.20 per million input tokens | |
| Replicate | meta/llama-4-maverick-instruct | $0.25/百万输入 | $0.25 per million input tokens | |
| OpenRouter | Meta: Llama 4 Maverick | $0.35/百万输入 | prompt: 0.00000035 / token | |
| OpenRouter | Meta: Llama 3.3 70B Instruct | $0.59/百万输入 | prompt: 0.00000059 / token | |
| Replicate | meta/llama-2-70b | $0.65/百万输入 | $0.65 per million input tokens | |
| Replicate | meta/llama-2-70b-chat | $0.65/百万输入 | $0.65 per million input tokens | |
| Replicate | meta/meta-llama-3-70b | $0.65/百万输入 | $0.65 per million input tokens | |
| Replicate | meta/meta-llama-3-70b-instruct | $0.65/百万输入 | $0.65 per million input tokens | |
| OpenRouter | Meta: Llama 3.1 70B Instruct | $0.72/百万输入 | prompt: 0.00000072 / token | |
| OpenRouter | AionLabs: Aion-RP 1.0 (8B) | $0.8/百万输入 | prompt: 0.0000008 / token | |
| OpenRouter | Sao10K: Llama 3.1 Euryale 70B v2.2 | $0.85/百万输入 | prompt: 0.00000085 / token |
每百万输出 token24
| 平台 | 模型 · 计价项 | 美元 | 平台原文 | 生效于 |
|---|---|---|---|---|
| OpenRouter | Sao10K: Llama 3 8B Lunaris | $0.05/百万输出最低 | completion: 0.00000005 / token | |
| OpenRouter | Meta: Llama 3.1 8B Instruct | $0.08/百万输出 | completion: 0.00000008 / token | |
| OpenRouter | Meta: Llama Guard 4 12B | $0.18/百万输出 | completion: 0.00000018 / token | |
| Replicate | meta/llama-guard-4-12b | $0.2/百万输出 | $0.20 per million output tokens | |
| OpenRouter | Meta: Llama 3.2 1B Instruct | $0.201/百万输出 | completion: 0.000000201 / token | |
| Replicate | meta/llama-2-7b | $0.25/百万输出 | $0.25 per million output tokens | |
| Replicate | meta/llama-2-7b-chat | $0.25/百万输出 | $0.25 per million output tokens | |
| Replicate | meta/meta-llama-3-8b | $0.25/百万输出 | $0.25 per million output tokens | |
| Replicate | meta/meta-llama-3-8b-instruct | $0.25/百万输出 | $0.25 per million output tokens | |
| OpenRouter | Meta: Llama 4 Scout | $0.3/百万输出 | completion: 0.0000003 / token | |
| OpenRouter | Meta: Llama 3.2 3B Instruct | $0.33/百万输出 | completion: 0.00000033 / token | |
| Replicate | meta/llama-2-13b | $0.5/百万输出 | $0.50 per million output tokens | |
| Replicate | meta/llama-2-13b-chat | $0.5/百万输出 | $0.50 per million output tokens | |
| Replicate | meta/llama-4-scout-instruct | $0.65/百万输出 | $0.65 per million output tokens | |
| OpenRouter | Meta: Llama 3.1 70B Instruct | $0.72/百万输出 | completion: 0.00000072 / token | |
| OpenRouter | Meta: Llama 3.3 70B Instruct | $0.79/百万输出 | completion: 0.00000079 / token | |
| OpenRouter | Sao10K: Llama 3.1 Euryale 70B v2.2 | $0.85/百万输出 | completion: 0.00000085 / token | |
| Replicate | meta/llama-4-maverick-instruct | $0.95/百万输出 | $0.95 per million output tokens | |
| OpenRouter | Meta: Llama 4 Maverick | $1/百万输出 | completion: 0.000001 / token | |
| OpenRouter | AionLabs: Aion-RP 1.0 (8B) | $1.6/百万输出 | completion: 0.0000016 / token | |
| Replicate | meta/llama-2-70b | $2.75/百万输出 | $2.75 per million output tokens | |
| Replicate | meta/llama-2-70b-chat | $2.75/百万输出 | $2.75 per million output tokens | |
| Replicate | meta/meta-llama-3-70b | $2.75/百万输出 | $2.75 per million output tokens | |
| Replicate | meta/meta-llama-3-70b-instruct | $2.75/百万输出 | $2.75 per million output tokens |
每百万缓存读取 token3
| 平台 | 模型 · 计价项 | 美元 | 平台原文 | 生效于 |
|---|---|---|---|---|
| OpenRouter | Meta: Llama 3.1 8B Instruct | $0.025/百万缓存读最低 | input_cache_read: 0.000000025 / token | |
| OpenRouter | Meta: Llama 4 Maverick | $0.17/百万缓存读 | input_cache_read: 0.00000017 / token | |
| OpenRouter | Meta: Llama 3.3 70B Instruct | $0.295/百万缓存读 | input_cache_read: 0.000000295 / token |