From 78f1cbefaa36f2d3f6444935a10db7316d16bc31 Mon Sep 17 00:00:00 2001 From: vov Date: Fri, 6 Jun 2025 03:54:08 +0300 Subject: [PATCH] feat(vertexai): update Gemini model versions and add new preview models - Add support for latest Gemini 2.5 Pro Preview (06-05) model - Add support for Gemini 2.5 Flash Preview (05-20) model - Add support for older preview models (05-06, 04-17) for backward compatibility - Update default model selections in config to use the latest preview versions - Replace generic model IDs with version-specific ones for better tracking --- internal/config/config.go | 12 ++++----- internal/llm/models/vertexai.go | 46 ++++++++++++++++++++++++++------- 2 files changed, 43 insertions(+), 15 deletions(-) diff --git a/internal/config/config.go b/internal/config/config.go index 5a0905bb..6410493e 100644 --- a/internal/config/config.go +++ b/internal/config/config.go @@ -352,10 +352,10 @@ func setProviderDefaults() { // Google Cloud VertexAI configuration if hasVertexAICredentials() { - viper.SetDefault("agents.coder.model", models.VertexAIGemini25) - viper.SetDefault("agents.summarizer.model", models.VertexAIGemini25) - viper.SetDefault("agents.task.model", models.VertexAIGemini25Flash) - viper.SetDefault("agents.title.model", models.VertexAIGemini25Flash) + viper.SetDefault("agents.coder.model", models.VertexAIGemini25ProPreview0605) + viper.SetDefault("agents.summarizer.model", models.VertexAIGemini25ProPreview0605) + viper.SetDefault("agents.task.model", models.VertexAIGemini25FlashPreview520) + viper.SetDefault("agents.title.model", models.VertexAIGemini25FlashPreview520) return } } @@ -750,10 +750,10 @@ func setDefaultModelForAgent(agent AgentName) bool { maxTokens := int64(5000) if agent == AgentTitle { - model = models.VertexAIGemini25Flash + model = models.VertexAIGemini25FlashPreview520 maxTokens = 80 } else { - model = models.VertexAIGemini25 + model = models.VertexAIGemini25ProPreview0605 } cfg.Agents[agent] = Agent{ diff --git a/internal/llm/models/vertexai.go b/internal/llm/models/vertexai.go index d71dfc0b..22b23d18 100644 --- a/internal/llm/models/vertexai.go +++ b/internal/llm/models/vertexai.go @@ -4,14 +4,29 @@ const ( ProviderVertexAI ModelProvider = "vertexai" // Models - VertexAIGemini25Flash ModelID = "vertexai.gemini-2.5-flash" - VertexAIGemini25 ModelID = "vertexai.gemini-2.5" + VertexAIGemini25FlashPreview520 ModelID = "vertexai.gemini-2.5-flash-preview-05-20" + VertexAIGemini25FlashPreview417 ModelID = "vertexai.gemini-2.5-flash-preview-04-17" + VertexAIGemini25ProPreview0605 ModelID = "vertexai.gemini-2.5-pro-preview-06-05" + VertexAIGemini25ProPreview0506 ModelID = "vertexai.gemini-2.5-pro-preview-05-06" ) var VertexAIGeminiModels = map[ModelID]Model{ - VertexAIGemini25Flash: { - ID: VertexAIGemini25Flash, - Name: "VertexAI: Gemini 2.5 Flash", + VertexAIGemini25FlashPreview520: { + ID: VertexAIGemini25FlashPreview520, + Name: "VertexAI: Gemini 2.5 Flash Preview (05-20)", + Provider: ProviderVertexAI, + APIModel: "gemini-2.5-flash-preview-05-20", + CostPer1MIn: GeminiModels[Gemini25Flash].CostPer1MIn, + CostPer1MInCached: GeminiModels[Gemini25Flash].CostPer1MInCached, + CostPer1MOut: GeminiModels[Gemini25Flash].CostPer1MOut, + CostPer1MOutCached: GeminiModels[Gemini25Flash].CostPer1MOutCached, + ContextWindow: GeminiModels[Gemini25Flash].ContextWindow, + DefaultMaxTokens: GeminiModels[Gemini25Flash].DefaultMaxTokens, + SupportsAttachments: true, + }, + VertexAIGemini25FlashPreview417: { + ID: VertexAIGemini25FlashPreview417, + Name: "VertexAI: Gemini 2.5 Flash Preview (04-17)", Provider: ProviderVertexAI, APIModel: "gemini-2.5-flash-preview-04-17", CostPer1MIn: GeminiModels[Gemini25Flash].CostPer1MIn, @@ -22,11 +37,24 @@ var VertexAIGeminiModels = map[ModelID]Model{ DefaultMaxTokens: GeminiModels[Gemini25Flash].DefaultMaxTokens, SupportsAttachments: true, }, - VertexAIGemini25: { - ID: VertexAIGemini25, - Name: "VertexAI: Gemini 2.5 Pro", + VertexAIGemini25ProPreview0605: { + ID: VertexAIGemini25ProPreview0605, + Name: "VertexAI: Gemini 2.5 Pro Preview (06-05)", + Provider: ProviderVertexAI, + APIModel: "gemini-2.5-pro-preview-06-05", + CostPer1MIn: GeminiModels[Gemini25].CostPer1MIn, + CostPer1MInCached: GeminiModels[Gemini25].CostPer1MInCached, + CostPer1MOut: GeminiModels[Gemini25].CostPer1MOut, + CostPer1MOutCached: GeminiModels[Gemini25].CostPer1MOutCached, + ContextWindow: GeminiModels[Gemini25].ContextWindow, + DefaultMaxTokens: GeminiModels[Gemini25].DefaultMaxTokens, + SupportsAttachments: true, + }, + VertexAIGemini25ProPreview0506: { + ID: VertexAIGemini25ProPreview0506, + Name: "VertexAI: Gemini 2.5 Pro Preview (05-06)", Provider: ProviderVertexAI, - APIModel: "gemini-2.5-pro-preview-03-25", + APIModel: "gemini-2.5-pro-preview-05-06", CostPer1MIn: GeminiModels[Gemini25].CostPer1MIn, CostPer1MInCached: GeminiModels[Gemini25].CostPer1MInCached, CostPer1MOut: GeminiModels[Gemini25].CostPer1MOut,