diff --git a/common/model-ratio.go b/common/model-ratio.go index 8b66f49..b84a0a5 100644 --- a/common/model-ratio.go +++ b/common/model-ratio.go @@ -69,9 +69,17 @@ var DefaultModelRatio = map[string]float64{ "claude-3-haiku-20240307": 0.125, // $0.25 / 1M tokens "claude-3-sonnet-20240229": 1.5, // $3 / 1M tokens "claude-3-opus-20240229": 7.5, // $15 / 1M tokens - "ERNIE-Bot": 0.8572, // ¥0.012 / 1k tokens - "ERNIE-Bot-turbo": 0.5715, // ¥0.008 / 1k tokens - "ERNIE-Bot-4": 8.572, // ¥0.12 / 1k tokens + "ERNIE-Bot": 0.8572, // ¥0.012 / 1k tokens //renamed to ERNIE-3.5-8K + "ERNIE-Bot-turbo": 0.5715, // ¥0.008 / 1k tokens //renamed to ERNIE-Lite-8K + "ERNIE-Bot-4": 8.572, // ¥0.12 / 1k tokens //renamed to ERNIE-4.0-8K + "ERNIE-4.0-8K": 8.572, // ¥0.12 / 1k tokens + "ERNIE-3.5-8K": 0.8572, // ¥0.012 / 1k tokens + "ERNIE-Speed-8K": 0.2858, // ¥0.004 / 1k tokens + "ERNIE-Speed-128K": 0.2858, // ¥0.004 / 1k tokens + "ERNIE-Lite-8K": 0.2143, // ¥0.003 / 1k tokens + "ERNIE-Tiny-8K": 0.0715, // ¥0.001 / 1k tokens + "ERNIE-Character-8K": 0.2858, // ¥0.004 / 1k tokens + "ERNIE-Functions-8K": 0.2858, // ¥0.004 / 1k tokens "Embedding-V1": 0.1429, // ¥0.002 / 1k tokens "PaLM-2": 1, "gemini-pro": 1, // $0.00025 / 1k characters -> $0.001 / 1k tokens @@ -98,6 +106,9 @@ var DefaultModelRatio = map[string]float64{ "SparkDesk-v3.1": 1.2858, // ¥0.018 / 1k tokens "SparkDesk-v3.5": 1.2858, // ¥0.018 / 1k tokens "360GPT_S2_V9": 0.8572, // ¥0.012 / 1k tokens + "360gpt-turbo": 0.0858, // ¥0.0012 / 1k tokens + "360gpt-turbo-responsibility-8k": 0.8572, // ¥0.012 / 1k tokens + "360gpt-pro": 0.8572, // ¥0.012 / 1k tokens "embedding-bert-512-v1": 0.0715, // ¥0.001 / 1k tokens "embedding_s1_v1": 0.0715, // ¥0.001 / 1k tokens "semantic_similarity_s1_v1": 0.0715, // ¥0.001 / 1k tokens @@ -289,6 +300,15 @@ func GetCompletionRatio(name string) float64 { if strings.HasPrefix(name, "deepseek") { return 2 } + if strings.HasPrefix(name, "ERNIE-Speed-") { + return 2 + } else if strings.HasPrefix(name, "ERNIE-Lite-") { + return 2 + } else if strings.HasPrefix(name, "ERNIE-Character") { + return 2 + } else if strings.HasPrefix(name, "ERNIE-Functions") { + return 2 + } switch name { case "llama2-70b-4096": return 0.8 / 0.64 diff --git a/relay/channel/ai360/constants.go b/relay/channel/ai360/constants.go index 82698fa..824231d 100644 --- a/relay/channel/ai360/constants.go +++ b/relay/channel/ai360/constants.go @@ -1,6 +1,9 @@ package ai360 var ModelList = []string{ + "360gpt-turbo", + "360gpt-turbo-responsibility-8k", + "360gpt-pro", "360GPT_S2_V9", "embedding-bert-512-v1", "embedding_s1_v1", diff --git a/relay/channel/baidu/adaptor.go b/relay/channel/baidu/adaptor.go index d2571dc..44c5e3f 100644 --- a/relay/channel/baidu/adaptor.go +++ b/relay/channel/baidu/adaptor.go @@ -9,6 +9,7 @@ import ( "one-api/relay/channel" relaycommon "one-api/relay/common" "one-api/relay/constant" + "strings" ) type Adaptor struct { @@ -33,8 +34,24 @@ func (a *Adaptor) GetRequestURL(info *relaycommon.RelayInfo) (string, error) { fullRequestURL = "https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/eb-instant" case "BLOOMZ-7B": fullRequestURL = "https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/bloomz_7b1" + case "ERNIE-4.0-8K": + fullRequestURL = "https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/completions_pro" + case "ERNIE-3.5-8K": + fullRequestURL = "https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/completions" + case "ERNIE-Speed-8K": + fullRequestURL = "https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/ernie_speed" + case "ERNIE-Character-8K": + fullRequestURL = "https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/ernie-char-8k" + case "ERNIE-Functions-8K": + fullRequestURL = "https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/ernie-func-8k" + case "ERNIE-Lite-8K-0922": + fullRequestURL = "https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/eb-instant" + case "Yi-34B-Chat": + fullRequestURL = "https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/yi_34b_chat" case "Embedding-V1": fullRequestURL = "https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/embeddings/embedding-v1" + default: + fullRequestURL = "https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/" + strings.ToLower(info.UpstreamModelName) } var accessToken string var err error diff --git a/relay/channel/baidu/constants.go b/relay/channel/baidu/constants.go index a0162bb..3cb96fc 100644 --- a/relay/channel/baidu/constants.go +++ b/relay/channel/baidu/constants.go @@ -1,11 +1,19 @@ package baidu var ModelList = []string{ - "ERNIE-Bot-4", - "ERNIE-Bot-8K", - "ERNIE-Bot", - "ERNIE-Speed", - "ERNIE-Bot-turbo", + "ERNIE-3.5-8K", + "ERNIE-4.0-8K", + "ERNIE-Speed-8K", + "ERNIE-Speed-128K", + "ERNIE-Lite-8K", + "ERNIE-Tiny-8K", + "ERNIE-Character-8K", + "ERNIE-Functions-8K", + //"ERNIE-Bot-4", + //"ERNIE-Bot-8K", + //"ERNIE-Bot", + //"ERNIE-Speed", + //"ERNIE-Bot-turbo", "Embedding-V1", }