feat: refactor token estimation logic

- Introduced new OpenAI text models in `common/model.go`.
- Added `IsOpenAITextModel` function to check for OpenAI text models.
- Refactored token estimation methods across various channels to use estimated prompt tokens instead of direct prompt token counts.
- Updated related functions and structures to accommodate the new token estimation approach, enhancing overall token management.

relay/channel/gemini/relay-gemini.go

@@ -1115,7 +1115,7 @@ func geminiStreamHandler(c *gin.Context, info *relaycommon.RelayInfo, resp *http
 	if usage.CompletionTokens <= 0 {
 		str := responseText.String()
 		if len(str) > 0 {
-			usage = service.ResponseText2Usage(c, responseText.String(), info.UpstreamModelName, info.PromptTokens)
+			usage = service.ResponseText2Usage(c, responseText.String(), info.UpstreamModelName, info.GetEstimatePromptTokens())
 		} else {
 			usage = &dto.Usage{}
 		}
@@ -1288,11 +1288,7 @@ func GeminiEmbeddingHandler(c *gin.Context, info *relaycommon.RelayInfo, resp *h
 	// Google has not yet clarified how embedding models will be billed
 	// refer to openai billing method to use input tokens billing
 	// https://platform.openai.com/docs/guides/embeddings#what-are-embeddings
-	usage := &dto.Usage{
-		PromptTokens:     info.PromptTokens,
-		CompletionTokens: 0,
-		TotalTokens:      info.PromptTokens,
-	}
+	usage := service.ResponseText2Usage(c, "", info.UpstreamModelName, info.GetEstimatePromptTokens())
 	openAIResponse.Usage = *usage
 
 	jsonResponse, jsonErr := common.Marshal(openAIResponse)