Skip to content

Open source AI coding agent. Designed for large projects and real world tasks.

wip: litellm/models refactor

Dane Schneiderda60aeb32f05
diff --git a/app/shared/ai_models_available.go b/app/shared/ai_models_available.go
index f0f6fefc..1df879ff 100644
--- a/app/shared/ai_models_available.go
+++ b/app/shared/ai_models_available.go
@@ -44,21 +44,17 @@ When checking for sufficient credits on Plandex Cloud, we use MaxOutputTokens-In
'ApiKeyEnvVar' is the environment variable that contains the API key for the model.
*/
-var AvailableModels = []*AvailableModel{
- // Direct OpenAI models
+var BuiltInModels = []*BaseModelConfigSchema{
{
- Description: "OpenAI o3-high",
+ ModelTag: "openai/o3",
+ Description: "OpenAI o3",
DefaultMaxConvoTokens: 15000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenAI,
- ModelName: "o3",
- ModelId: "openai/o3-high",
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 200000,
MaxOutputTokens: 100000,
ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
- ApiKeyEnvVar: OpenAIEnvVar,
ModelCompatibility: fullCompatibility,
- BaseUrl: OpenAIV1BaseUrl,
PreferredModelOutputFormat: ModelOutputFormatXml,
SystemPromptDisabled: true,
RoleParamsDisabled: true,
@@ -66,21 +62,40 @@ var AvailableModels = []*AvailableModel{
ReasoningEffort: ReasoningEffortHigh,
StopDisabled: true,
},
- },
+ Variants: []BaseModelConfigVariant{
+ {
+ VariantTag: "high",
+ Description: "high",
+ },
+ },
+
+ Providers: []BaseModelUsesProvider{
+ {
+ Provider: ModelProviderOpenAI,
+ ModelName: "o3",
+ },
+ {
+ Provider: ModelProviderAzureOpenAI,
+ ModelName: "o3",
+ },
+ {
+ Provider: ModelProviderOpenRouter,
+ ModelName: "openai/o3",
+ },
+ },
+ },
{
+ PriceId: "openai/o3",
+ ModelTag: "openai/o3-medium",
Description: "OpenAI o3-medium",
DefaultMaxConvoTokens: 15000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenAI,
- ModelName: "o3",
- ModelId: "openai/o3-medium",
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 200000,
MaxOutputTokens: 100000,
ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
- ApiKeyEnvVar: OpenAIEnvVar,
ModelCompatibility: fullCompatibility,
- BaseUrl: OpenAIV1BaseUrl,
PreferredModelOutputFormat: ModelOutputFormatXml,
SystemPromptDisabled: true,
RoleParamsDisabled: true,
@@ -88,21 +103,33 @@ var AvailableModels = []*AvailableModel{
ReasoningEffort: ReasoningEffortMedium,
StopDisabled: true,
},
- },
+ Providers: []BaseModelUsesProvider{
+ {
+ Provider: ModelProviderOpenAI,
+ ModelName: "o3",
+ },
+ {
+ Provider: ModelProviderAzureOpenAI,
+ ModelName: "o3",
+ },
+ {
+ Provider: ModelProviderOpenRouter,
+ ModelName: "openai/o3",
+ },
+ },
+ },
{
+ PriceId: "openai/o3",
+ ModelTag: "openai/o3-low",
Description: "OpenAI o3-low",
DefaultMaxConvoTokens: 15000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenAI,
- ModelName: "o3",
- ModelId: "openai/o3-low",
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 200000,
MaxOutputTokens: 100000,
ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
- ApiKeyEnvVar: OpenAIEnvVar,
ModelCompatibility: fullCompatibility,
- BaseUrl: OpenAIV1BaseUrl,
PreferredModelOutputFormat: ModelOutputFormatXml,
SystemPromptDisabled: true,
RoleParamsDisabled: true,
@@ -110,741 +137,1545 @@ var AvailableModels = []*AvailableModel{
ReasoningEffort: ReasoningEffortLow,
StopDisabled: true,
},
+
+ Providers: []BaseModelUsesProvider{
+ {
+ Provider: ModelProviderOpenAI,
+ ModelName: "o3",
+ },
+ {
+ Provider: ModelProviderAzureOpenAI,
+ ModelName: "o3",
+ },
+ {
+ Provider: ModelProviderOpenRouter,
+ ModelName: "openai/o3",
+ },
+ },
},
{
+ PriceId: "openai/o4-mini",
+ ModelTag: "openai/o4-mini-high",
Description: "OpenAI o4-mini-high",
DefaultMaxConvoTokens: 10000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenAI,
- ModelName: "o4-mini",
- ModelId: "openai/o4-mini-high",
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 200000,
MaxOutputTokens: 100000,
- ReservedOutputTokens: 30000,
- ApiKeyEnvVar: OpenAIEnvVar,
+ ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
ModelCompatibility: fullCompatibility,
- BaseUrl: OpenAIV1BaseUrl,
PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+ SystemPromptDisabled: true,
RoleParamsDisabled: true,
ReasoningEffortEnabled: true,
ReasoningEffort: ReasoningEffortHigh,
StopDisabled: true,
},
+
+ Providers: []BaseModelUsesProvider{
+ {
+ Provider: ModelProviderOpenAI,
+ ModelName: "o4-mini",
+ },
+ {
+ Provider: ModelProviderAzureOpenAI,
+ ModelName: "o4-mini",
+ },
+ {
+ Provider: ModelProviderOpenRouter,
+ ModelName: "openai/o4-mini",
+ },
+ },
},
+
{
+ PriceId: "openai/o4-mini",
+ ModelTag: "openai/o4-mini-medium",
Description: "OpenAI o4-mini-medium",
DefaultMaxConvoTokens: 10000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenAI,
- ModelName: "o4-mini",
- ModelId: "openai/o4-mini-medium",
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 200000,
MaxOutputTokens: 100000,
ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
- ApiKeyEnvVar: OpenAIEnvVar,
ModelCompatibility: fullCompatibility,
- BaseUrl: OpenAIV1BaseUrl,
PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+ SystemPromptDisabled: true,
RoleParamsDisabled: true,
ReasoningEffortEnabled: true,
ReasoningEffort: ReasoningEffortMedium,
StopDisabled: true,
},
+
+ Providers: []BaseModelUsesProvider{
+ {
+ Provider: ModelProviderOpenAI,
+ ModelName: "o4-mini",
+ },
+ {
+ Provider: ModelProviderAzureOpenAI,
+ ModelName: "o4-mini",
+ },
+ {
+ Provider: ModelProviderOpenRouter,
+ ModelName: "openai/o4-mini",
+ },
+ },
},
+
{
+ PriceId: "openai/o4-mini",
+ ModelTag: "openai/o4-mini-low",
Description: "OpenAI o4-mini-low",
DefaultMaxConvoTokens: 10000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenAI,
- ModelName: "o4-mini",
- ModelId: "openai/o4-mini-low",
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 200000,
MaxOutputTokens: 100000,
ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
- ApiKeyEnvVar: OpenAIEnvVar,
ModelCompatibility: fullCompatibility,
- BaseUrl: OpenAIV1BaseUrl,
PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+ SystemPromptDisabled: true,
RoleParamsDisabled: true,
ReasoningEffortEnabled: true,
ReasoningEffort: ReasoningEffortLow,
StopDisabled: true,
},
+
+ Providers: []BaseModelUsesProvider{
+ {
+ Provider: ModelProviderOpenAI,
+ ModelName: "o4-mini",
+ },
+ {
+ Provider: ModelProviderAzureOpenAI,
+ ModelName: "o4-mini",
+ },
+ {
+ Provider: ModelProviderOpenRouter,
+ ModelName: "openai/o4-mini",
+ },
+ },
},
{
+ PriceId: "openai/gpt-4.1",
+ ModelTag: "openai/gpt-4.1",
Description: "OpenAI gpt-4.1",
DefaultMaxConvoTokens: 75000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenAI,
- ModelName: "gpt-4.1",
- ModelId: "openai/gpt-4.1",
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 1047576,
MaxOutputTokens: 32768,
ReservedOutputTokens: 32768,
- ApiKeyEnvVar: OpenAIEnvVar,
ModelCompatibility: fullCompatibility,
- BaseUrl: OpenAIV1BaseUrl,
PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
},
+
+ Providers: []BaseModelUsesProvider{
+ {
+ Provider: ModelProviderOpenAI,
+ ModelName: "gpt-4.1",
+ },
+ {
+ Provider: ModelProviderAzureOpenAI,
+ ModelName: "gpt-4.1",
+ },
+ {
+ Provider: ModelProviderOpenRouter,
+ ModelName: "openai/gpt-4.1",
+ },
+ },
},
{
+ PriceId: "openai/gpt-4.1-mini",
+ ModelTag: "openai/gpt-4.1-mini",
Description: "OpenAI gpt-4.1-mini",
DefaultMaxConvoTokens: 75000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenAI,
- ModelName: "gpt-4.1-mini",
- ModelId: "openai/gpt-4.1-mini",
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 1047576,
MaxOutputTokens: 32768,
ReservedOutputTokens: 32768,
- ApiKeyEnvVar: OpenAIEnvVar,
ModelCompatibility: fullCompatibility,
- BaseUrl: OpenAIV1BaseUrl,
PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
},
+
+ Providers: []BaseModelUsesProvider{
+ {
+ Provider: ModelProviderOpenAI,
+ ModelName: "gpt-4.1-mini",
+ },
+ {
+ Provider: ModelProviderAzureOpenAI,
+ ModelName: "gpt-4.1-mini",
+ },
+ {
+ Provider: ModelProviderOpenRouter,
+ ModelName: "openai/gpt-4.1-mini",
+ },
+ },
},
{
+ PriceId: "openai/gpt-4.1-nano",
+ ModelTag: "openai/gpt-4.1-nano",
Description: "OpenAI gpt-4.1-nano",
DefaultMaxConvoTokens: 75000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenAI,
- ModelName: "gpt-4.1-nano",
- ModelId: "openai/gpt-4.1-nano",
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 1047576,
MaxOutputTokens: 32768,
ReservedOutputTokens: 32768,
- ApiKeyEnvVar: OpenAIEnvVar,
ModelCompatibility: fullCompatibility,
- BaseUrl: OpenAIV1BaseUrl,
PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
},
+
+ Providers: []BaseModelUsesProvider{
+ {
+ Provider: ModelProviderOpenAI,
+ ModelName: "gpt-4.1-nano",
+ },
+ {
+ Provider: ModelProviderAzureOpenAI,
+ ModelName: "gpt-4.1-nano",
+ },
+ {
+ Provider: ModelProviderOpenRouter,
+ ModelName: "openai/gpt-4.1-nano",
+ },
+ },
},
{
+ PriceId: "openai/o3-mini",
+ ModelTag: "openai/o3-mini-high",
Description: "OpenAI o3-mini-high",
DefaultMaxConvoTokens: 10000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenAI,
- ModelName: "o3-mini",
- ModelId: "openai/o3-mini-high",
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 200000,
MaxOutputTokens: 100000,
- ReservedOutputTokens: 30000,
- ApiKeyEnvVar: OpenAIEnvVar,
+ ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
ModelCompatibility: fullCompatibility,
- BaseUrl: OpenAIV1BaseUrl,
PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+ SystemPromptDisabled: true,
RoleParamsDisabled: true,
ReasoningEffortEnabled: true,
ReasoningEffort: ReasoningEffortHigh,
+ StopDisabled: true,
+ },
+
+ Providers: []BaseModelUsesProvider{
+ {
+ Provider: ModelProviderOpenAI,
+ ModelName: "o3-mini",
+ },
+ {
+ Provider: ModelProviderAzureOpenAI,
+ ModelName: "o3-mini",
+ },
+ {
+ Provider: ModelProviderOpenRouter,
+ ModelName: "openai/o3-mini",
+ },
},
},
+
{
+ PriceId: "openai/o3-mini",
+ ModelTag: "openai/o3-mini-medium",
Description: "OpenAI o3-mini-medium",
DefaultMaxConvoTokens: 10000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenAI,
- ModelName: "o3-mini",
- ModelId: "openai/o3-mini-medium",
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 200000,
MaxOutputTokens: 100000,
ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
- ApiKeyEnvVar: OpenAIEnvVar,
ModelCompatibility: fullCompatibility,
- BaseUrl: OpenAIV1BaseUrl,
PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+ SystemPromptDisabled: true,
RoleParamsDisabled: true,
ReasoningEffortEnabled: true,
ReasoningEffort: ReasoningEffortMedium,
+ StopDisabled: true,
+ },
+
+ Providers: []BaseModelUsesProvider{
+ {
+ Provider: ModelProviderOpenAI,
+ ModelName: "o3-mini",
+ },
+ {
+ Provider: ModelProviderAzureOpenAI,
+ ModelName: "o3-mini",
+ },
+ {
+ Provider: ModelProviderOpenRouter,
+ ModelName: "openai/o3-mini",
+ },
},
},
+
{
+ PriceId: "openai/o3-mini",
+ ModelTag: "openai/o3-mini-low",
Description: "OpenAI o3-mini-low",
DefaultMaxConvoTokens: 10000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenAI,
- ModelName: "o3-mini",
- ModelId: "openai/o3-mini-low",
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 200000,
MaxOutputTokens: 100000,
ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
- ApiKeyEnvVar: OpenAIEnvVar,
ModelCompatibility: fullCompatibility,
- BaseUrl: OpenAIV1BaseUrl,
PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+ SystemPromptDisabled: true,
RoleParamsDisabled: true,
ReasoningEffortEnabled: true,
ReasoningEffort: ReasoningEffortLow,
+ StopDisabled: true,
+ },
+
+ Providers: []BaseModelUsesProvider{
+ {
+ Provider: ModelProviderOpenAI,
+ ModelName: "o3-mini",
+ },
+ {
+ Provider: ModelProviderAzureOpenAI,
+ ModelName: "o3-mini",
+ },
+ {
+ Provider: ModelProviderOpenRouter,
+ ModelName: "openai/o3-mini",
+ },
},
},
- // OpenRouter models
{
- Description: "Anthropic Claude 3.7 Sonnet via OpenRouter",
+ PriceId: "anthropic/claude-3.7-sonnet",
+ ModelTag: "anthropic/claude-3.7-sonnet",
+ Description: "Anthropic Claude 3.7 Sonnet",
DefaultMaxConvoTokens: 15000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "anthropic/claude-3.7-sonnet",
- ModelId: "anthropic/claude-3.7-sonnet",
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 200000,
MaxOutputTokens: 128000,
ReservedOutputTokens: 20000,
SupportsCacheControl: true,
- ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
PreferredModelOutputFormat: ModelOutputFormatXml,
SingleMessageNoSystemPrompt: true,
TokenEstimatePaddingPct: 0.10,
},
+
+ Providers: []BaseModelUsesProvider{
+ {
+ Provider: ModelProviderAnthropic,
+ ModelName: "claude-3.7-sonnet",
+ },
+ {
+ Provider: ModelProviderGoogleVertex,
+ ModelName: "claude-3-7-sonnet@20250219",
+ },
+ {
+ Provider: ModelProviderAmazonBedrock,
+ ModelName: "anthropic.claude-3-7-sonnet-20250219-v1:0",
+ },
+ {
+ Provider: ModelProviderOpenRouter,
+ ModelName: "anthropic/claude-3.7-sonnet",
+ },
+ },
},
+
{
- Description: "Anthropic Claude 3.7 Sonnet (thinking—includes reasoning) via OpenRouter",
+ PriceId: "anthropic/claude-3.7-sonnet",
+ ModelTag: "anthropic/claude-3.7-sonnet:thinking-hidden",
+ Description: "Anthropic Claude 3.7 Sonnet (thinking—reasoning hidden) ",
DefaultMaxConvoTokens: 15000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "anthropic/claude-3.7-sonnet:thinking",
- ModelId: "anthropic/claude-3.7-sonnet:thinking",
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 200000,
MaxOutputTokens: 128000,
- ReservedOutputTokens: 40000,
+ ReservedOutputTokens: 20000,
SupportsCacheControl: true,
- ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
PreferredModelOutputFormat: ModelOutputFormatXml,
- IncludeReasoning: true,
SingleMessageNoSystemPrompt: true,
TokenEstimatePaddingPct: 0.10,
+ IncludeReasoning: false,
+ ReasoningBudget: AnthropicMaxReasoningBudget,
+ },
+
+ Providers: []BaseModelUsesProvider{
+ {
+ Provider: ModelProviderAnthropic,
+ ModelName: "claude-3.7-sonnet",
+ },
+ {
+ Provider: ModelProviderGoogleVertex,
+ ModelName: "claude-3-7-sonnet@20250219",
+ },
+ {
+ Provider: ModelProviderAmazonBedrock,
+ ModelName: "anthropic.claude-3-7-sonnet-20250219-v1:0",
+ },
+ {
+ Provider: ModelProviderOpenRouter,
+ ModelName: "anthropic/claude-3.7-sonnet",
+ },
},
},
+
{
- Description: "Anthropic Claude 3.7 Sonnet (thinking—reasoning hidden) via OpenRouter",
+ PriceId: "anthropic/claude-3.7-sonnet",
+ ModelTag: "anthropic/claude-3.7-sonnet:thinking-visible",
+ Description: "Anthropic Claude 3.7 Sonnet (thinking—reasoning visible) ",
DefaultMaxConvoTokens: 15000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "anthropic/claude-3.7-sonnet:thinking",
- ModelId: "anthropic/claude-3.7-sonnet:thinking-hidden",
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 200000,
MaxOutputTokens: 128000,
- ReservedOutputTokens: 40000,
+ ReservedOutputTokens: 20000,
SupportsCacheControl: true,
- ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
PreferredModelOutputFormat: ModelOutputFormatXml,
- IncludeReasoning: false,
SingleMessageNoSystemPrompt: true,
TokenEstimatePaddingPct: 0.10,
+ IncludeReasoning: true,
+ ReasoningBudget: AnthropicMaxReasoningBudget,
},
- },
- {
- Description: "Anthropic Claude 3.7 Sonnet",
- DefaultMaxConvoTokens: 15000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderAnthropic,
- ModelName: "claude-3-7-sonnet-latest",
- ModelId: "claude-3-7-sonnet-latest",
- MaxTokens: 200000,
- MaxOutputTokens: 128000,
- ReservedOutputTokens: 20000,
- SupportsCacheControl: true,
- ApiKeyEnvVar: ApiKeyByProvider[ModelProviderAnthropic],
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderAnthropic],
- PreferredModelOutputFormat: ModelOutputFormatXml,
+
+ Providers: []BaseModelUsesProvider{
+ {
+ Provider: ModelProviderAnthropic,
+ ModelName: "claude-3.7-sonnet",
+ },
+ {
+ Provider: ModelProviderGoogleVertex,
+ ModelName: "claude-3-7-sonnet@20250219",
+ },
+ {
+ Provider: ModelProviderAmazonBedrock,
+ ModelName: "anthropic.claude-3-7-sonnet-20250219-v1:0",
+ },
+ {
+ Provider: ModelProviderOpenRouter,
+ ModelName: "anthropic/claude-3.7-sonnet",
+ },
},
},
+
{
- Description: "Anthropic Claude 3.5 Sonnet via OpenRouter",
+ PriceId: "anthropic/claude-3.5-sonnet",
+ ModelTag: "anthropic/claude-3.5-sonnet",
+ Description: "Anthropic Claude 3.5 Sonnet",
DefaultMaxConvoTokens: 15000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "anthropic/claude-3.5-sonnet",
- ModelId: "anthropic/claude-3.5-sonnet",
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 200000,
MaxOutputTokens: 128000,
ReservedOutputTokens: 20000,
SupportsCacheControl: true,
- ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
PreferredModelOutputFormat: ModelOutputFormatXml,
SingleMessageNoSystemPrompt: true,
TokenEstimatePaddingPct: 0.10,
},
+
+ Providers: []BaseModelUsesProvider{
+ {
+ Provider: ModelProviderAnthropic,
+ ModelName: "claude-3.5-sonnet",
+ },
+ {
+ Provider: ModelProviderGoogleVertex,
+ ModelName: "claude-3-5-sonnet@20240620",
+ },
+ {
+ Provider: ModelProviderAmazonBedrock,
+ ModelName: "anthropic.claude-3-5-sonnet-20241022-v2:0",
+ },
+ {
+ Provider: ModelProviderOpenRouter,
+ ModelName: "anthropic/claude-3.5-sonnet",
+ },
+ },
},
+
{
- Description: "Anthropic Claude 3.5 Haiku via OpenRouter",
+ PriceId: "anthropic/claude-3.5-haiku",
+ ModelTag: "anthropic/claude-3.5-haiku",
+ Description: "Anthropic Claude 3.5 Haiku",
DefaultMaxConvoTokens: 15000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "anthropic/claude-3.5-haiku",
- ModelId: "anthropic/claude-3.5-haiku",
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 200000,
MaxOutputTokens: 8192,
ReservedOutputTokens: 8192,
SupportsCacheControl: true,
- ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
PreferredModelOutputFormat: ModelOutputFormatXml,
SingleMessageNoSystemPrompt: true,
TokenEstimatePaddingPct: 0.10,
},
+
+ Providers: []BaseModelUsesProvider{
+ {
+ Provider: ModelProviderAnthropic,
+ ModelName: "claude-3.5-haiku",
+ },
+ {
+ Provider: ModelProviderGoogleVertex,
+ ModelName: "claude-3-5-haiku@20241022",
+ },
+ {
+ Provider: ModelProviderAmazonBedrock,
+ ModelName: "anthropic.claude-3-5-haiku-20241022-v1:0",
+ },
+ {
+ Provider: ModelProviderOpenRouter,
+ ModelName: "anthropic/claude-3.5-haiku",
+ },
+ },
},
+
{
- Description: "Google Gemini Pro 1.5 via OpenRouter",
- DefaultMaxConvoTokens: 100000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "google/gemini-pro-1.5",
- ModelId: "google/gemini-pro-1.5",
+ PriceId: "google/gemini-pro-1.5",
+ ModelTag: "google/gemini-pro-1.5",
+ Description: "Google Gemini 1.5 Pro",
+ DefaultMaxConvoTokens: 75000,
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 2000000,
MaxOutputTokens: 8192,
ReservedOutputTokens: 8192,
- ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
- PreferredModelOutputFormat: ModelOutputFormatXml,
SupportsCacheControl: true,
+ PreferredModelOutputFormat: ModelOutputFormatXml,
},
- },
+ Providers: []BaseModelUsesProvider{
+ {Provider: ModelProviderGoogleAIStudio, ModelName: "models/gemini-1.5-pro"},
+ {Provider: ModelProviderGoogleVertex, ModelName: "gemini-1.5-pro-latest"},
+ {Provider: ModelProviderOpenRouter, ModelName: "google/gemini-pro-1.5"},
+ },
+ },
{
- Description: "Google Gemini Pro 2.5 Preview via OpenRouter",
+ PriceId: "google/gemini-2.5-pro-preview",
+ ModelTag: "google/gemini-2.5-pro-preview",
+ Description: "Google Gemini 2.5 Pro (Preview)",
DefaultMaxConvoTokens: 75000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "google/gemini-2.5-pro-preview-03-25",
- ModelId: "google/gemini-2.5-pro-preview-03-25",
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 1048576,
MaxOutputTokens: 65535,
ReservedOutputTokens: 65535,
- ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
- PreferredModelOutputFormat: ModelOutputFormatXml,
SupportsCacheControl: true,
+ PreferredModelOutputFormat: ModelOutputFormatXml,
},
- },
+ Providers: []BaseModelUsesProvider{
+ {Provider: ModelProviderGoogleAIStudio, ModelName: "models/gemini-2.5-pro-preview-03-25"},
+ {Provider: ModelProviderGoogleVertex, ModelName: "gemini-2.5-pro-preview-03-25"},
+ {Provider: ModelProviderOpenRouter, ModelName: "google/gemini-2.5-pro-preview-03-25"},
+ },
+ },
{
- Description: "Google Gemini Pro 2.5 Experimental via OpenRouter",
+ PriceId: "google/gemini-2.5-pro-exp",
+ ModelTag: "google/gemini-2.5-pro-exp",
+ Description: "Google Gemini 2.5 Pro (Experimental)",
DefaultMaxConvoTokens: 75000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "google/gemini-2.5-pro-exp-03-25",
- ModelId: "google/gemini-2.5-pro-exp-03-25",
- MaxTokens: 1000000,
+
+ BaseModelShared: BaseModelShared{
+ MaxTokens: 1048576,
MaxOutputTokens: 65535,
ReservedOutputTokens: 65535,
- ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
PreferredModelOutputFormat: ModelOutputFormatXml,
},
- },
- {
- Description: "Google Gemini Pro 1.5 Pro via AI Studio",
- DefaultMaxConvoTokens: 75000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderGoogleAIStudio,
- ModelName: "models/gemini-1.5-pro",
- ModelId: "models/gemini-1.5-pro",
- MaxTokens: 2000000,
- MaxOutputTokens: 8192,
- ReservedOutputTokens: 8192,
- ApiKeyEnvVar: GoogleAIStudioApiKeyEnvVar,
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderGoogleAIStudio],
- PreferredModelOutputFormat: ModelOutputFormatXml,
+ Providers: []BaseModelUsesProvider{
+ {Provider: ModelProviderGoogleAIStudio, ModelName: "models/gemini-2.5-pro-exp-03-25"},
+ {Provider: ModelProviderGoogleVertex, ModelName: "gemini-2.5-pro-exp-03-25"},
+ {Provider: ModelProviderOpenRouter, ModelName: "google/gemini-2.5-pro-exp-03-25"},
},
},
-
{
- Description: "Google Gemini Flash 2.5 Preview via OpenRouter",
+ PriceId: "google/gemini-2.5-flash-preview",
+ ModelTag: "google/gemini-2.5-flash-preview",
+ Description: "Google Gemini 2.5 Flash (Preview)",
DefaultMaxConvoTokens: 75000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "google/gemini-2.5-flash-preview",
- ModelId: "google/gemini-2.5-flash-preview",
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 1048576,
MaxOutputTokens: 65535,
ReservedOutputTokens: 65535,
- ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
- PreferredModelOutputFormat: ModelOutputFormatXml,
SupportsCacheControl: true,
+ PreferredModelOutputFormat: ModelOutputFormatXml,
+ },
+
+ Providers: []BaseModelUsesProvider{
+ {Provider: ModelProviderGoogleAIStudio, ModelName: "models/gemini-2.5-flash-preview-04-17"},
+ {Provider: ModelProviderGoogleVertex, ModelName: "gemini-2.5-flash-preview-04-17"},
+ {Provider: ModelProviderOpenRouter, ModelName: "google/gemini-2.5-flash-preview"},
},
},
{
- Description: "DeepSeek V3 0324 via OpenRouter",
+ PriceId: "deepseek/v3-0324",
+ ModelTag: "deepseek/v3-0324",
+ Description: "DeepSeek V3 (0324)",
DefaultMaxConvoTokens: 7500,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "deepseek/deepseek-chat-v3-0324",
- ModelId: "deepseek/deepseek-chat-v3-0324",
- MaxTokens: 64000,
- MaxOutputTokens: 8192,
- ReservedOutputTokens: 8192,
- ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
- ModelCompatibility: ModelCompatibility{
- HasImageSupport: false,
- },
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+
+ BaseModelShared: BaseModelShared{
+ MaxTokens: 64000,
+ MaxOutputTokens: 8192,
+ ReservedOutputTokens: 8192,
PreferredModelOutputFormat: ModelOutputFormatXml,
},
- },
+ Providers: []BaseModelUsesProvider{
+ {Provider: ModelProviderDeepSeek, ModelName: "deepseek-chat"},
+ {Provider: ModelProviderOpenRouter, ModelName: "deepseek/deepseek-chat-v3-0324"},
+ },
+ },
{
- Description: "DeepSeek R1 via OpenRouter (includes reasoning)",
+ PriceId: "deepseek/r1",
+ ModelTag: "deepseek/r1",
+ Description: "DeepSeek R1 (includes reasoning)",
DefaultMaxConvoTokens: 7500,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "deepseek/deepseek-r1",
- ModelId: "deepseek/deepseek-r1-reasoning",
- MaxTokens: 64000,
- MaxOutputTokens: 8192,
- ReservedOutputTokens: 8192,
- ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
- ModelCompatibility: ModelCompatibility{
- HasImageSupport: false,
- },
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+
+ BaseModelShared: BaseModelShared{
+ MaxTokens: 64000,
+ MaxOutputTokens: 8192,
+ ReservedOutputTokens: 8192,
PreferredModelOutputFormat: ModelOutputFormatXml,
IncludeReasoning: true,
},
- },
+ Providers: []BaseModelUsesProvider{
+ {Provider: ModelProviderDeepSeek, ModelName: "deepseek-reasoner"},
+ {Provider: ModelProviderOpenRouter, ModelName: "deepseek/deepseek-r1"},
+ },
+ },
{
- Description: "DeepSeek R1 via OpenRouter (reasoning hidden)",
+ PriceId: "deepseek/r1",
+ ModelTag: "deepseek/r1:reasoning-hidden",
+ Description: "DeepSeek R1 (reasoning hidden)",
DefaultMaxConvoTokens: 7500,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "deepseek/deepseek-r1",
- ModelId: "deepseek/deepseek-r1-no-reasoning",
- MaxTokens: 64000,
- MaxOutputTokens: 8192,
- ReservedOutputTokens: 8192,
- ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
- ModelCompatibility: ModelCompatibility{
- HasImageSupport: false,
- },
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+
+ BaseModelShared: BaseModelShared{
+ MaxTokens: 64000,
+ MaxOutputTokens: 8192,
+ ReservedOutputTokens: 8192,
PreferredModelOutputFormat: ModelOutputFormatXml,
+ IncludeReasoning: false,
+ },
+
+ Providers: []BaseModelUsesProvider{
+ {Provider: ModelProviderDeepSeek, ModelName: "deepseek-reasoner"},
+ {Provider: ModelProviderOpenRouter, ModelName: "deepseek/deepseek-r1"},
},
},
{
- Description: "Perplexity R1 1776 via OpenRouter (includes reasoning)",
+ PriceId: "perplexity/r1-1776",
+ ModelTag: "perplexity/r1-1776",
+ Description: "Perplexity R1-1776 (includes reasoning)",
DefaultMaxConvoTokens: 7500,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "perplexity/r1-1776",
- ModelId: "perplexity/r1-1776",
- MaxTokens: 128000,
- MaxOutputTokens: 128000,
- ReservedOutputTokens: 30000,
- ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
- ModelCompatibility: ModelCompatibility{
- HasImageSupport: false,
- },
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+
+ BaseModelShared: BaseModelShared{
+ MaxTokens: 128000,
+ MaxOutputTokens: 128000,
+ ReservedOutputTokens: 30000,
PreferredModelOutputFormat: ModelOutputFormatXml,
IncludeReasoning: true,
},
- },
+ Providers: []BaseModelUsesProvider{
+ {Provider: ModelProviderPerplexity, ModelName: "r1-1776-online"},
+ {Provider: ModelProviderOpenRouter, ModelName: "perplexity/r1-1776"},
+ },
+ },
{
- Description: "Perplexity Sonar Reasoning via OpenRouter (includes reasoning)",
+ PriceId: "perplexity/sonar-reasoning",
+ ModelTag: "perplexity/sonar-reasoning",
+ Description: "Perplexity Sonar Reasoning (includes reasoning)",
DefaultMaxConvoTokens: 7500,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "perplexity/sonar-reasoning",
- ModelId: "perplexity/sonar-reasoning",
- MaxTokens: 127000,
- MaxOutputTokens: 127000,
- ReservedOutputTokens: 30000,
- ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
- ModelCompatibility: ModelCompatibility{
- HasImageSupport: false,
- },
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+
+ BaseModelShared: BaseModelShared{
+ MaxTokens: 127000,
+ MaxOutputTokens: 127000,
+ ReservedOutputTokens: 30000,
PreferredModelOutputFormat: ModelOutputFormatXml,
IncludeReasoning: true,
},
+
+ Providers: []BaseModelUsesProvider{
+ {Provider: ModelProviderPerplexity, ModelName: "sonar-reasoning-online"},
+ {Provider: ModelProviderOpenRouter, ModelName: "perplexity/sonar-reasoning"},
+ },
},
{
- Description: "Qwen 2.5 Coder 32B via OpenRouter",
+ PriceId: "qwen/qwen-2.5-coder-32b-instruct",
+ ModelTag: "qwen/qwen-2.5-coder-32b-instruct",
+ Description: "Qwen 2.5 Coder 32B (Instruct)",
DefaultMaxConvoTokens: 10000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "qwen/qwen-2.5-coder-32b-instruct",
- ModelId: "qwen/qwen-2.5-coder-32b-instruct",
+
+ BaseModelShared: BaseModelShared{
MaxTokens: 128000,
MaxOutputTokens: 8192,
ReservedOutputTokens: 8192,
- ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
PreferredModelOutputFormat: ModelOutputFormatXml,
},
- },
-
- // OpenAI models via OpenRouter
- {
- Description: "OpenAI o3-high via OpenRouter",
- DefaultMaxConvoTokens: 15000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "openai/o3",
- ModelId: "openai/o3-high",
- MaxTokens: 200000,
- MaxOutputTokens: 100000,
- ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
- ApiKeyEnvVar: OpenRouterApiKeyEnvVar,
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
- PreferredModelOutputFormat: ModelOutputFormatXml,
- SystemPromptDisabled: true,
- RoleParamsDisabled: true,
- ReasoningEffortEnabled: true,
- ReasoningEffort: ReasoningEffortHigh,
- StopDisabled: true,
+ Providers: []BaseModelUsesProvider{
+ {Provider: ModelProviderOpenRouter, ModelName: "qwen/qwen-2.5-coder-32b-instruct"},
},
},
{
- Description: "OpenAI o3-medium via OpenRouter",
- DefaultMaxConvoTokens: 15000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "openai/o3",
- ModelId: "openai/o3-medium",
- MaxTokens: 200000,
- MaxOutputTokens: 100000,
- ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
- ApiKeyEnvVar: OpenRouterApiKeyEnvVar,
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
- PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
- SystemPromptDisabled: true,
- RoleParamsDisabled: true,
- ReasoningEffortEnabled: true,
- ReasoningEffort: ReasoningEffortMedium,
- StopDisabled: true,
- },
- },
+ PriceId: "qwen/qwen3-32b",
+ ModelTag: "qwen/qwen3-32b",
+ Description: "Qwen 3-32B (Experimental)",
+ DefaultMaxConvoTokens: 15000, // feel free to raise once you test
- {
- Description: "OpenAI o3-low via OpenRouter",
- DefaultMaxConvoTokens: 15000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "openai/o3",
- ModelId: "openai/o3-low",
- MaxTokens: 200000,
- MaxOutputTokens: 100000,
- ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
- ApiKeyEnvVar: OpenRouterApiKeyEnvVar,
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+ BaseModelShared: BaseModelShared{
+ MaxTokens: 128000,
+ MaxOutputTokens: 16384,
+ ReservedOutputTokens: 16384,
PreferredModelOutputFormat: ModelOutputFormatXml,
- SystemPromptDisabled: true,
- RoleParamsDisabled: true,
- ReasoningEffortEnabled: true,
- ReasoningEffort: ReasoningEffortLow,
- StopDisabled: true,
+ // leave IncludeReasoning unset – Qwen3 has its own “thinking” mode but OR
+ // exposes it via a special prompt key, not a flag
},
- },
- {
- Description: "OpenAI o4-mini-high via OpenRouter",
- DefaultMaxConvoTokens: 10000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "openai/o4-mini",
- ModelId: "openai/o4-mini-high",
- MaxTokens: 200000,
- MaxOutputTokens: 100000,
- ReservedOutputTokens: 30000,
- ApiKeyEnvVar: OpenRouterApiKeyEnvVar,
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
- PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
- RoleParamsDisabled: true,
- ReasoningEffortEnabled: true,
- ReasoningEffort: ReasoningEffortHigh,
- StopDisabled: true,
+ Providers: []BaseModelUsesProvider{
+ {Provider: ModelProviderOpenRouter, ModelName: "qwen/qwen3-32b-a3b"}, // OR cloud
+ // optional local provider once you wire up Ollama or HF Inference
+ // {Provider: ModelProviderOllama, ModelName: "qwen3-32b"},
},
},
{
- Description: "OpenAI o4-mini-medium via OpenRouter",
- DefaultMaxConvoTokens: 10000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "openai/o4-mini",
- ModelId: "openai/o4-mini-medium",
- MaxTokens: 200000,
- MaxOutputTokens: 100000,
- ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
- ApiKeyEnvVar: OpenRouterApiKeyEnvVar,
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
- PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
- RoleParamsDisabled: true,
- ReasoningEffortEnabled: true,
- ReasoningEffort: ReasoningEffortMedium,
- StopDisabled: true,
- },
- },
- {
- Description: "OpenAI o4-mini-low via OpenRouter",
- DefaultMaxConvoTokens: 10000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "openai/o4-mini",
- ModelId: "openai/o4-mini-low",
- MaxTokens: 200000,
- MaxOutputTokens: 100000,
- ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
- ApiKeyEnvVar: OpenRouterApiKeyEnvVar,
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
- PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
- RoleParamsDisabled: true,
- ReasoningEffortEnabled: true,
- ReasoningEffort: ReasoningEffortLow,
- StopDisabled: true,
- },
- },
+ PriceId: "qwen/qwen3-8b",
+ ModelTag: "qwen/qwen3-8b",
+ Description: "Qwen 3-8B (Experimental)",
+ DefaultMaxConvoTokens: 7500,
- {
- Description: "OpenAI gpt-4.1 via OpenRouter",
- DefaultMaxConvoTokens: 75000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "openai/gpt-4.1",
- ModelId: "openai/gpt-4.1",
- MaxTokens: 1047576,
- MaxOutputTokens: 32768,
- ReservedOutputTokens: 32768,
- ApiKeyEnvVar: OpenRouterApiKeyEnvVar,
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
- PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+ BaseModelShared: BaseModelShared{
+ MaxTokens: 32768,
+ MaxOutputTokens: 8192,
+ ReservedOutputTokens: 8192,
+ PreferredModelOutputFormat: ModelOutputFormatXml,
},
- },
- {
- Description: "OpenAI gpt-4.1-mini via OpenRouter",
- DefaultMaxConvoTokens: 75000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "openai/gpt-4.1-mini",
- ModelId: "openai/gpt-4.1-mini",
- MaxTokens: 1047576,
- MaxOutputTokens: 32768,
- ReservedOutputTokens: 32768,
- ApiKeyEnvVar: OpenRouterApiKeyEnvVar,
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
- PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+ Providers: []BaseModelUsesProvider{
+ {Provider: ModelProviderOpenRouter, ModelName: "qwen/qwen3-8b-04-28"},
},
},
+}
- {
- Description: "OpenAI gpt-4.1-nano via OpenRouter",
- DefaultMaxConvoTokens: 75000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "openai/gpt-4.1-nano",
- ModelId: "openai/gpt-4.1-nano",
- MaxTokens: 1047576,
- MaxOutputTokens: 32768,
- ReservedOutputTokens: 32768,
- ApiKeyEnvVar: OpenRouterApiKeyEnvVar,
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
- PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
- },
- },
+var AvailableModels = []*AvailableModel{}
- {
- Description: "OpenAI o3-mini-high via OpenRouter",
- DefaultMaxConvoTokens: 10000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "openai/o3-mini",
- ModelId: "openai/o3-mini-high",
- MaxTokens: 200000,
- MaxOutputTokens: 100000,
- ReservedOutputTokens: 40000,
- ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
- PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
- SystemPromptDisabled: true,
- RoleParamsDisabled: true,
- ReasoningEffortEnabled: true,
- ReasoningEffort: ReasoningEffortHigh,
- },
- },
- {
- Description: "OpenAI o3-mini-medium via OpenRouter",
- DefaultMaxConvoTokens: 10000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "openai/o3-mini",
- ModelId: "openai/o3-mini-medium",
- MaxTokens: 200000,
- MaxOutputTokens: 100000,
- ReservedOutputTokens: 40000,
- ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
- PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
- SystemPromptDisabled: true,
- RoleParamsDisabled: true,
- ReasoningEffortEnabled: true,
- ReasoningEffort: ReasoningEffortMedium,
- },
- },
- {
- Description: "OpenAI o3-mini-low via OpenRouter",
- DefaultMaxConvoTokens: 10000,
- BaseModelConfig: BaseModelConfig{
- Provider: ModelProviderOpenRouter,
- ModelName: "openai/o3-mini",
- ModelId: "openai/o3-mini-low",
- MaxTokens: 200000,
- MaxOutputTokens: 100000,
- ReservedOutputTokens: 40000,
- ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
- ModelCompatibility: fullCompatibility,
- BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
- PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
- SystemPromptDisabled: true,
- RoleParamsDisabled: true,
- ReasoningEffortEnabled: true,
- ReasoningEffort: ReasoningEffortLow,
- },
- },
-}
+// var AvailableModels = []*AvailableModel{
+// // Direct OpenAI models
+// {
+// Description: "OpenAI o3-high",
+// DefaultMaxConvoTokens: 15000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenAI,
+// ModelName: "o3",
+// ModelId: "openai/o3-high",
+// MaxTokens: 200000,
+// MaxOutputTokens: 100000,
+// ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
+// ApiKeyEnvVar: OpenAIEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: OpenAIV1BaseUrl,
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// SystemPromptDisabled: true,
+// RoleParamsDisabled: true,
+// ReasoningEffortEnabled: true,
+// ReasoningEffort: ReasoningEffortHigh,
+// StopDisabled: true,
+// },
+// },
+
+// {
+// Description: "OpenAI o3-medium",
+// DefaultMaxConvoTokens: 15000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenAI,
+// ModelName: "o3",
+// ModelId: "openai/o3-medium",
+// MaxTokens: 200000,
+// MaxOutputTokens: 100000,
+// ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
+// ApiKeyEnvVar: OpenAIEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: OpenAIV1BaseUrl,
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// SystemPromptDisabled: true,
+// RoleParamsDisabled: true,
+// ReasoningEffortEnabled: true,
+// ReasoningEffort: ReasoningEffortMedium,
+// StopDisabled: true,
+// },
+// },
+
+// {
+// Description: "OpenAI o3-low",
+// DefaultMaxConvoTokens: 15000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenAI,
+// ModelName: "o3",
+// ModelId: "openai/o3-low",
+// MaxTokens: 200000,
+// MaxOutputTokens: 100000,
+// ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
+// ApiKeyEnvVar: OpenAIEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: OpenAIV1BaseUrl,
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// SystemPromptDisabled: true,
+// RoleParamsDisabled: true,
+// ReasoningEffortEnabled: true,
+// ReasoningEffort: ReasoningEffortLow,
+// StopDisabled: true,
+// },
+// },
+
+// {
+// Description: "OpenAI o4-mini-high",
+// DefaultMaxConvoTokens: 10000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenAI,
+// ModelName: "o4-mini",
+// ModelId: "openai/o4-mini-high",
+// MaxTokens: 200000,
+// MaxOutputTokens: 100000,
+// ReservedOutputTokens: 30000,
+// ApiKeyEnvVar: OpenAIEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: OpenAIV1BaseUrl,
+// PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+// RoleParamsDisabled: true,
+// ReasoningEffortEnabled: true,
+// ReasoningEffort: ReasoningEffortHigh,
+// StopDisabled: true,
+// },
+// },
+// {
+// Description: "OpenAI o4-mini-medium",
+// DefaultMaxConvoTokens: 10000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenAI,
+// ModelName: "o4-mini",
+// ModelId: "openai/o4-mini-medium",
+// MaxTokens: 200000,
+// MaxOutputTokens: 100000,
+// ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
+// ApiKeyEnvVar: OpenAIEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: OpenAIV1BaseUrl,
+// PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+// RoleParamsDisabled: true,
+// ReasoningEffortEnabled: true,
+// ReasoningEffort: ReasoningEffortMedium,
+// StopDisabled: true,
+// },
+// },
+// {
+// Description: "OpenAI o4-mini-low",
+// DefaultMaxConvoTokens: 10000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenAI,
+// ModelName: "o4-mini",
+// ModelId: "openai/o4-mini-low",
+// MaxTokens: 200000,
+// MaxOutputTokens: 100000,
+// ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
+// ApiKeyEnvVar: OpenAIEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: OpenAIV1BaseUrl,
+// PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+// RoleParamsDisabled: true,
+// ReasoningEffortEnabled: true,
+// ReasoningEffort: ReasoningEffortLow,
+// StopDisabled: true,
+// },
+// },
+
+// {
+// Description: "OpenAI gpt-4.1",
+// DefaultMaxConvoTokens: 75000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenAI,
+// ModelName: "gpt-4.1",
+// ModelId: "openai/gpt-4.1",
+// MaxTokens: 1047576,
+// MaxOutputTokens: 32768,
+// ReservedOutputTokens: 32768,
+// ApiKeyEnvVar: OpenAIEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: OpenAIV1BaseUrl,
+// PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+// },
+// },
+
+// {
+// Description: "OpenAI gpt-4.1-mini",
+// DefaultMaxConvoTokens: 75000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenAI,
+// ModelName: "gpt-4.1-mini",
+// ModelId: "openai/gpt-4.1-mini",
+// MaxTokens: 1047576,
+// MaxOutputTokens: 32768,
+// ReservedOutputTokens: 32768,
+// ApiKeyEnvVar: OpenAIEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: OpenAIV1BaseUrl,
+// PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+// },
+// },
+
+// {
+// Description: "OpenAI gpt-4.1-nano",
+// DefaultMaxConvoTokens: 75000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenAI,
+// ModelName: "gpt-4.1-nano",
+// ModelId: "openai/gpt-4.1-nano",
+// MaxTokens: 1047576,
+// MaxOutputTokens: 32768,
+// ReservedOutputTokens: 32768,
+// ApiKeyEnvVar: OpenAIEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: OpenAIV1BaseUrl,
+// PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+// },
+// },
+
+// {
+// Description: "OpenAI o3-mini-high",
+// DefaultMaxConvoTokens: 10000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenAI,
+// ModelName: "o3-mini",
+// ModelId: "openai/o3-mini-high",
+// MaxTokens: 200000,
+// MaxOutputTokens: 100000,
+// ReservedOutputTokens: 30000,
+// ApiKeyEnvVar: OpenAIEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: OpenAIV1BaseUrl,
+// PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+// RoleParamsDisabled: true,
+// ReasoningEffortEnabled: true,
+// ReasoningEffort: ReasoningEffortHigh,
+// },
+// },
+// {
+// Description: "OpenAI o3-mini-medium",
+// DefaultMaxConvoTokens: 10000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenAI,
+// ModelName: "o3-mini",
+// ModelId: "openai/o3-mini-medium",
+// MaxTokens: 200000,
+// MaxOutputTokens: 100000,
+// ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
+// ApiKeyEnvVar: OpenAIEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: OpenAIV1BaseUrl,
+// PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+// RoleParamsDisabled: true,
+// ReasoningEffortEnabled: true,
+// ReasoningEffort: ReasoningEffortMedium,
+// },
+// },
+// {
+// Description: "OpenAI o3-mini-low",
+// DefaultMaxConvoTokens: 10000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenAI,
+// ModelName: "o3-mini",
+// ModelId: "openai/o3-mini-low",
+// MaxTokens: 200000,
+// MaxOutputTokens: 100000,
+// ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
+// ApiKeyEnvVar: OpenAIEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: OpenAIV1BaseUrl,
+// PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+// RoleParamsDisabled: true,
+// ReasoningEffortEnabled: true,
+// ReasoningEffort: ReasoningEffortLow,
+// },
+// },
+
+// // OpenRouter models
+// {
+// Description: "Anthropic Claude 3.7 Sonnet via OpenRouter",
+// DefaultMaxConvoTokens: 15000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "anthropic/claude-3.7-sonnet",
+// ModelId: "anthropic/claude-3.7-sonnet",
+// MaxTokens: 200000,
+// MaxOutputTokens: 128000,
+// ReservedOutputTokens: 20000,
+// SupportsCacheControl: true,
+// ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// SingleMessageNoSystemPrompt: true,
+// TokenEstimatePaddingPct: 0.10,
+// },
+// },
+// {
+// Description: "Anthropic Claude 3.7 Sonnet (thinking—includes reasoning) via OpenRouter",
+// DefaultMaxConvoTokens: 15000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "anthropic/claude-3.7-sonnet:thinking",
+// ModelId: "anthropic/claude-3.7-sonnet:thinking",
+// MaxTokens: 200000,
+// MaxOutputTokens: 128000,
+// ReservedOutputTokens: 40000,
+// SupportsCacheControl: true,
+// ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// IncludeReasoning: true,
+// SingleMessageNoSystemPrompt: true,
+// TokenEstimatePaddingPct: 0.10,
+// },
+// },
+// {
+// Description: "Anthropic Claude 3.7 Sonnet (thinking—reasoning hidden) via OpenRouter",
+// DefaultMaxConvoTokens: 15000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "anthropic/claude-3.7-sonnet:thinking",
+// ModelId: "anthropic/claude-3.7-sonnet:thinking-hidden",
+// MaxTokens: 200000,
+// MaxOutputTokens: 128000,
+// ReservedOutputTokens: 40000,
+// SupportsCacheControl: true,
+// ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// IncludeReasoning: false,
+// SingleMessageNoSystemPrompt: true,
+// TokenEstimatePaddingPct: 0.10,
+// },
+// },
+// {
+// Description: "Anthropic Claude 3.7 Sonnet",
+// DefaultMaxConvoTokens: 15000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderAnthropic,
+// ModelName: "claude-3-7-sonnet-latest",
+// ModelId: "claude-3-7-sonnet-latest",
+// MaxTokens: 200000,
+// MaxOutputTokens: 128000,
+// ReservedOutputTokens: 20000,
+// SupportsCacheControl: true,
+// ApiKeyEnvVar: ApiKeyByProvider[ModelProviderAnthropic],
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderAnthropic],
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// },
+// },
+// {
+// Description: "Anthropic Claude 3.5 Sonnet via OpenRouter",
+// DefaultMaxConvoTokens: 15000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "anthropic/claude-3.5-sonnet",
+// ModelId: "anthropic/claude-3.5-sonnet",
+// MaxTokens: 200000,
+// MaxOutputTokens: 128000,
+// ReservedOutputTokens: 20000,
+// SupportsCacheControl: true,
+// ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// SingleMessageNoSystemPrompt: true,
+// TokenEstimatePaddingPct: 0.10,
+// },
+// },
+// {
+// Description: "Anthropic Claude 3.5 Haiku via OpenRouter",
+// DefaultMaxConvoTokens: 15000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "anthropic/claude-3.5-haiku",
+// ModelId: "anthropic/claude-3.5-haiku",
+// MaxTokens: 200000,
+// MaxOutputTokens: 8192,
+// ReservedOutputTokens: 8192,
+// SupportsCacheControl: true,
+// ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// SingleMessageNoSystemPrompt: true,
+// TokenEstimatePaddingPct: 0.10,
+// },
+// },
+// {
+// Description: "Google Gemini Pro 1.5 via OpenRouter",
+// DefaultMaxConvoTokens: 100000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "google/gemini-pro-1.5",
+// ModelId: "google/gemini-pro-1.5",
+// MaxTokens: 2000000,
+// MaxOutputTokens: 8192,
+// ReservedOutputTokens: 8192,
+// ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// SupportsCacheControl: true,
+// },
+// },
+
+// {
+// Description: "Google Gemini Pro 2.5 Preview via OpenRouter",
+// DefaultMaxConvoTokens: 75000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "google/gemini-2.5-pro-preview-03-25",
+// ModelId: "google/gemini-2.5-pro-preview-03-25",
+// MaxTokens: 1048576,
+// MaxOutputTokens: 65535,
+// ReservedOutputTokens: 65535,
+// ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// SupportsCacheControl: true,
+// },
+// },
+
+// {
+// Description: "Google Gemini Pro 2.5 Experimental via OpenRouter",
+// DefaultMaxConvoTokens: 75000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "google/gemini-2.5-pro-exp-03-25",
+// ModelId: "google/gemini-2.5-pro-exp-03-25",
+// MaxTokens: 1000000,
+// MaxOutputTokens: 65535,
+// ReservedOutputTokens: 65535,
+// ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// },
+// },
+
+// {
+// Description: "Google Gemini Pro 1.5 Pro via AI Studio",
+// DefaultMaxConvoTokens: 75000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderGoogleAIStudio,
+// ModelName: "models/gemini-1.5-pro",
+// ModelId: "models/gemini-1.5-pro",
+// MaxTokens: 2000000,
+// MaxOutputTokens: 8192,
+// ReservedOutputTokens: 8192,
+// ApiKeyEnvVar: GoogleAIStudioApiKeyEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderGoogleAIStudio],
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// },
+// },
+
+// {
+// Description: "Google Gemini Flash 2.5 Preview via OpenRouter",
+// DefaultMaxConvoTokens: 75000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "google/gemini-2.5-flash-preview",
+// ModelId: "google/gemini-2.5-flash-preview",
+// MaxTokens: 1048576,
+// MaxOutputTokens: 65535,
+// ReservedOutputTokens: 65535,
+// ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// SupportsCacheControl: true,
+// },
+// },
+
+// {
+// Description: "DeepSeek V3 0324 via OpenRouter",
+// DefaultMaxConvoTokens: 7500,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "deepseek/deepseek-chat-v3-0324",
+// ModelId: "deepseek/deepseek-chat-v3-0324",
+// MaxTokens: 64000,
+// MaxOutputTokens: 8192,
+// ReservedOutputTokens: 8192,
+// ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
+// ModelCompatibility: ModelCompatibility{
+// HasImageSupport: false,
+// },
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// },
+// },
+
+// {
+// Description: "DeepSeek R1 via OpenRouter (includes reasoning)",
+// DefaultMaxConvoTokens: 7500,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "deepseek/deepseek-r1",
+// ModelId: "deepseek/deepseek-r1-reasoning",
+// MaxTokens: 64000,
+// MaxOutputTokens: 8192,
+// ReservedOutputTokens: 8192,
+// ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
+// ModelCompatibility: ModelCompatibility{
+// HasImageSupport: false,
+// },
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// IncludeReasoning: true,
+// },
+// },
+
+// {
+// Description: "DeepSeek R1 via OpenRouter (reasoning hidden)",
+// DefaultMaxConvoTokens: 7500,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "deepseek/deepseek-r1",
+// ModelId: "deepseek/deepseek-r1-no-reasoning",
+// MaxTokens: 64000,
+// MaxOutputTokens: 8192,
+// ReservedOutputTokens: 8192,
+// ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
+// ModelCompatibility: ModelCompatibility{
+// HasImageSupport: false,
+// },
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// },
+// },
+
+// {
+// Description: "Perplexity R1 1776 via OpenRouter (includes reasoning)",
+// DefaultMaxConvoTokens: 7500,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "perplexity/r1-1776",
+// ModelId: "perplexity/r1-1776",
+// MaxTokens: 128000,
+// MaxOutputTokens: 128000,
+// ReservedOutputTokens: 30000,
+// ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
+// ModelCompatibility: ModelCompatibility{
+// HasImageSupport: false,
+// },
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// IncludeReasoning: true,
+// },
+// },
+
+// {
+// Description: "Perplexity Sonar Reasoning via OpenRouter (includes reasoning)",
+// DefaultMaxConvoTokens: 7500,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "perplexity/sonar-reasoning",
+// ModelId: "perplexity/sonar-reasoning",
+// MaxTokens: 127000,
+// MaxOutputTokens: 127000,
+// ReservedOutputTokens: 30000,
+// ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
+// ModelCompatibility: ModelCompatibility{
+// HasImageSupport: false,
+// },
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// IncludeReasoning: true,
+// },
+// },
+
+// {
+// Description: "Qwen 2.5 Coder 32B via OpenRouter",
+// DefaultMaxConvoTokens: 10000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "qwen/qwen-2.5-coder-32b-instruct",
+// ModelId: "qwen/qwen-2.5-coder-32b-instruct",
+// MaxTokens: 128000,
+// MaxOutputTokens: 8192,
+// ReservedOutputTokens: 8192,
+// ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// },
+// },
+
+// // OpenAI models via OpenRouter
+
+// {
+// Description: "OpenAI o3-high via OpenRouter",
+// DefaultMaxConvoTokens: 15000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "openai/o3",
+// ModelId: "openai/o3-high",
+// MaxTokens: 200000,
+// MaxOutputTokens: 100000,
+// ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
+// ApiKeyEnvVar: OpenRouterApiKeyEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// SystemPromptDisabled: true,
+// RoleParamsDisabled: true,
+// ReasoningEffortEnabled: true,
+// ReasoningEffort: ReasoningEffortHigh,
+// StopDisabled: true,
+// },
+// },
+
+// {
+// Description: "OpenAI o3-medium via OpenRouter",
+// DefaultMaxConvoTokens: 15000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "openai/o3",
+// ModelId: "openai/o3-medium",
+// MaxTokens: 200000,
+// MaxOutputTokens: 100000,
+// ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
+// ApiKeyEnvVar: OpenRouterApiKeyEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+// SystemPromptDisabled: true,
+// RoleParamsDisabled: true,
+// ReasoningEffortEnabled: true,
+// ReasoningEffort: ReasoningEffortMedium,
+// StopDisabled: true,
+// },
+// },
+
+// {
+// Description: "OpenAI o3-low via OpenRouter",
+// DefaultMaxConvoTokens: 15000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "openai/o3",
+// ModelId: "openai/o3-low",
+// MaxTokens: 200000,
+// MaxOutputTokens: 100000,
+// ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
+// ApiKeyEnvVar: OpenRouterApiKeyEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatXml,
+// SystemPromptDisabled: true,
+// RoleParamsDisabled: true,
+// ReasoningEffortEnabled: true,
+// ReasoningEffort: ReasoningEffortLow,
+// StopDisabled: true,
+// },
+// },
+
+// {
+// Description: "OpenAI o4-mini-high via OpenRouter",
+// DefaultMaxConvoTokens: 10000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "openai/o4-mini",
+// ModelId: "openai/o4-mini-high",
+// MaxTokens: 200000,
+// MaxOutputTokens: 100000,
+// ReservedOutputTokens: 30000,
+// ApiKeyEnvVar: OpenRouterApiKeyEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+// RoleParamsDisabled: true,
+// ReasoningEffortEnabled: true,
+// ReasoningEffort: ReasoningEffortHigh,
+// StopDisabled: true,
+// },
+// },
+// {
+// Description: "OpenAI o4-mini-medium via OpenRouter",
+// DefaultMaxConvoTokens: 10000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "openai/o4-mini",
+// ModelId: "openai/o4-mini-medium",
+// MaxTokens: 200000,
+// MaxOutputTokens: 100000,
+// ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
+// ApiKeyEnvVar: OpenRouterApiKeyEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+// RoleParamsDisabled: true,
+// ReasoningEffortEnabled: true,
+// ReasoningEffort: ReasoningEffortMedium,
+// StopDisabled: true,
+// },
+// },
+// {
+// Description: "OpenAI o4-mini-low via OpenRouter",
+// DefaultMaxConvoTokens: 10000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "openai/o4-mini",
+// ModelId: "openai/o4-mini-low",
+// MaxTokens: 200000,
+// MaxOutputTokens: 100000,
+// ReservedOutputTokens: 40000, // 25k for reasoning, 15k for output
+// ApiKeyEnvVar: OpenRouterApiKeyEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+// RoleParamsDisabled: true,
+// ReasoningEffortEnabled: true,
+// ReasoningEffort: ReasoningEffortLow,
+// StopDisabled: true,
+// },
+// },
+
+// {
+// Description: "OpenAI gpt-4.1 via OpenRouter",
+// DefaultMaxConvoTokens: 75000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "openai/gpt-4.1",
+// ModelId: "openai/gpt-4.1",
+// MaxTokens: 1047576,
+// MaxOutputTokens: 32768,
+// ReservedOutputTokens: 32768,
+// ApiKeyEnvVar: OpenRouterApiKeyEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+// },
+// },
+
+// {
+// Description: "OpenAI gpt-4.1-mini via OpenRouter",
+// DefaultMaxConvoTokens: 75000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "openai/gpt-4.1-mini",
+// ModelId: "openai/gpt-4.1-mini",
+// MaxTokens: 1047576,
+// MaxOutputTokens: 32768,
+// ReservedOutputTokens: 32768,
+// ApiKeyEnvVar: OpenRouterApiKeyEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+// },
+// },
+
+// {
+// Description: "OpenAI gpt-4.1-nano via OpenRouter",
+// DefaultMaxConvoTokens: 75000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "openai/gpt-4.1-nano",
+// ModelId: "openai/gpt-4.1-nano",
+// MaxTokens: 1047576,
+// MaxOutputTokens: 32768,
+// ReservedOutputTokens: 32768,
+// ApiKeyEnvVar: OpenRouterApiKeyEnvVar,
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+// },
+// },
+
+// {
+// Description: "OpenAI o3-mini-high via OpenRouter",
+// DefaultMaxConvoTokens: 10000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "openai/o3-mini",
+// ModelId: "openai/o3-mini-high",
+// MaxTokens: 200000,
+// MaxOutputTokens: 100000,
+// ReservedOutputTokens: 40000,
+// ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+// SystemPromptDisabled: true,
+// RoleParamsDisabled: true,
+// ReasoningEffortEnabled: true,
+// ReasoningEffort: ReasoningEffortHigh,
+// },
+// },
+// {
+// Description: "OpenAI o3-mini-medium via OpenRouter",
+// DefaultMaxConvoTokens: 10000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "openai/o3-mini",
+// ModelId: "openai/o3-mini-medium",
+// MaxTokens: 200000,
+// MaxOutputTokens: 100000,
+// ReservedOutputTokens: 40000,
+// ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+// SystemPromptDisabled: true,
+// RoleParamsDisabled: true,
+// ReasoningEffortEnabled: true,
+// ReasoningEffort: ReasoningEffortMedium,
+// },
+// },
+// {
+// Description: "OpenAI o3-mini-low via OpenRouter",
+// DefaultMaxConvoTokens: 10000,
+// BaseModelConfig: BaseModelConfig{
+// Provider: ModelProviderOpenRouter,
+// ModelName: "openai/o3-mini",
+// ModelId: "openai/o3-mini-low",
+// MaxTokens: 200000,
+// MaxOutputTokens: 100000,
+// ReservedOutputTokens: 40000,
+// ApiKeyEnvVar: ApiKeyByProvider[ModelProviderOpenRouter],
+// ModelCompatibility: fullCompatibility,
+// BaseUrl: BaseUrlByProvider[ModelProviderOpenRouter],
+// PreferredModelOutputFormat: ModelOutputFormatToolCallJson,
+// SystemPromptDisabled: true,
+// RoleParamsDisabled: true,
+// ReasoningEffortEnabled: true,
+// ReasoningEffort: ReasoningEffortLow,
+// },
+// },
+// }
var AvailableModelsByComposite = map[string]*AvailableModel{}
func init() {
+
for _, model := range AvailableModels {
if model.Description == "" {
spew.Dump(model)
@@ -880,9 +1711,9 @@ func init() {
panic("reserved output tokens is not set")
}
- if model.ApiKeyEnvVar == "" {
+ if model.ApiKeyEnvVar == "" && len(model.ExtraAuthVars) == 0 && !model.SkipAuth && !model.HasAWSAuth {
spew.Dump(model)
- panic("api key env var is not set")
+ panic("api key or auth settings are not set")
}
if model.BaseUrl == "" {
diff --git a/app/shared/ai_models_data_models.go b/app/shared/ai_models_data_models.go
index a416118e..7ed95ee9 100644
--- a/app/shared/ai_models_data_models.go
+++ b/app/shared/ai_models_data_models.go
@@ -4,6 +4,7 @@ import (
"database/sql/driver"
"encoding/json"
"fmt"
+ "strings"
"time"
)
@@ -18,13 +19,13 @@ const (
ModelOutputFormatXml ModelOutputFormat = "xml"
)
-// to help avoid confusion between model tag, model name, and the model id
+// to help avoid confusion between model tag, price id, model name, and the model id
type ModelName string
type ModelId string
type ModelTag string
+type VariantTag string
type BaseModelShared struct {
- ModelTag ModelTag `json:"modelTag"`
MaxTokens int `json:"maxTokens"`
MaxOutputTokens int `json:"maxOutputTokens"`
ReservedOutputTokens int `json:"reservedOutputTokens"`
@@ -36,31 +37,73 @@ type BaseModelShared struct {
ReasoningEffortEnabled bool `json:"reasoningEffortEnabled"`
ReasoningEffort ReasoningEffort `json:"reasoningEffort"`
IncludeReasoning bool `json:"includeReasoning"`
+ ReasoningBudget int `json:"reasoningBudget"`
SupportsCacheControl bool `json:"supportsCacheControl"`
// for anthropic, single message system prompt needs to be flipped to 'user'
SingleMessageNoSystemPrompt bool `json:"singleMessageNoSystemPrompt"`
// for openai, token estimate padding percentage
TokenEstimatePaddingPct float64 `json:"tokenEstimatePaddingPct"`
+
+ ModelCompatibility ModelCompatibility `json:"modelCompatibility"`
}
type BaseModelProviderConfig struct {
- Provider ModelProvider `json:"provider"`
- CustomProvider *string `json:"customProvider,omitempty"`
- ModelName ModelName `json:"modelName"`
- ModelId ModelId `json:"modelId"`
+ ModelProviderConfigSchema
+ ModelName ModelName `json:"modelName"`
}
type BaseModelConfig struct {
- ApiKeyEnvVar string `json:"apiKeyEnvVar"`
+ ModelId ModelId `json:"modelId"`
BaseModelShared
BaseModelProviderConfig
}
+type BaseModelUsesProvider struct {
+ Provider ModelProvider `json:"provider"`
+ CustomProvider *string `json:"customProvider,omitempty"`
+ ModelName ModelName `json:"modelName"`
+}
+
type BaseModelConfigSchema struct {
- SchemaVersion string `json:"schemaVersion"`
+ ModelTag ModelTag `json:"modelTag"`
+ Description string `json:"description"`
+ DefaultMaxConvoTokens int `json:"defaultMaxConvoTokens"`
BaseModelShared
- Providers []BaseModelProviderConfig `json:"providers"`
+ Variants []BaseModelConfigVariant `json:"variants"`
+ Providers []BaseModelUsesProvider `json:"providers"`
+}
+
+type BaseModelConfigVariant struct {
+ VariantTag VariantTag `json:"variantTag"`
+ Description string `json:"description"`
+ BaseModelShared
+ Variants []BaseModelConfigVariant `json:"variants"`
+}
+
+func (b *BaseModelConfigSchema) ToAvailableModels() []*AvailableModel {
+ avail := []*AvailableModel{}
+ for _, provider := range b.Providers {
+
+ providerConfig, ok := BuiltInModelProviderConfigs[provider.Provider]
+ if !ok {
+ panic(fmt.Sprintf("provider %s not found", provider.Provider))
+ }
+
+ avail = append(avail, &AvailableModel{
+ Description: b.Description,
+ DefaultMaxConvoTokens: b.DefaultMaxConvoTokens,
+ BaseModelConfig: BaseModelConfig{
+ ModelId: ModelId(strings.Join([]string{string(provider.Provider), string(provider.ModelName)}, "/")),
+ BaseModelShared: b.BaseModelShared,
+ BaseModelProviderConfig: BaseModelProviderConfig{
+ ModelProviderConfigSchema: providerConfig,
+ ModelName: provider.ModelName,
+ },
+ },
+ })
+ }
+ return avail
}
type AvailableModel struct {
@@ -81,15 +124,6 @@ func (m *AvailableModel) ModelString() string {
return s
}
-type AvailableModelSchema struct {
- SchemaVersion string `json:"schemaVersion"`
- BaseModelConfigSchema
- Description string `json:"description"`
- DefaultMaxConvoTokens int `json:"defaultMaxConvoTokens"`
- CreatedAt time.Time `json:"createdAt"`
- UpdatedAt time.Time `json:"updatedAt"`
-}
-
type PlannerModelConfig struct {
MaxConvoTokens int `json:"maxConvoTokens"`
}
@@ -116,7 +150,77 @@ type ModelRoleConfig struct {
StrongModel *ModelRoleConfig `json:"strongModel"`
}
+type ModelRoleModelConfig struct {
+ Provider ModelProvider `json:"provider"`
+ CustomProvider *string `json:"customProvider,omitempty"`
+ ModelTag ModelTag `json:"modelTag"`
+}
+
type ModelRoleConfigSchema struct {
+ Role ModelRole `json:"role"`
+ ModelTag ModelTag `json:"modelTag"`
+
+ Temperature float32 `json:"temperature"`
+ TopP float32 `json:"topP"`
+ ReservedOutputTokens int `json:"reservedOutputTokens"`
+ MaxConvoTokens int `json:"maxConvoTokens"`
+
+ LargeContextFallback *ModelRoleConfigSchema `json:"largeContextFallback"`
+ LargeOutputFallback *ModelRoleConfigSchema `json:"largeOutputFallback"`
+ ErrorFallback *ModelRoleConfigSchema `json:"errorFallback"`
+ StrongModel *ModelRoleConfigSchema `json:"strongModel"`
+}
+
+func (m *ModelRoleConfigSchema) ToModelRoleConfig() ModelRoleConfig {
+
+ modelSchema
+
+ return m.toModelRoleConfig()
+}
+
+func (m *ModelRoleConfigSchema) toModelRoleConfig(providers []BaseModelUsesProvider) ModelRoleConfig {
+ if len(providers) == 0 {
+ panic("no providers")
+ }
+
+ provider := providers[0]
+ tail := providers[1:]
+
+ config := GetAvailableModel(m.Provider, m.ModelTag)
+
+ var largeContextFallback *ModelRoleConfig
+ if m.LargeContextFallback != nil {
+ c := m.LargeContextFallback.ToModelRoleConfig()
+ largeContextFallback = &c
+ }
+ var largeOutputFallback *ModelRoleConfig
+ if m.LargeOutputFallback != nil {
+ c := m.LargeOutputFallback.ToModelRoleConfig()
+ largeOutputFallback = &c
+ }
+ var errorFallback *ModelRoleConfig
+ if m.ErrorFallback != nil {
+ c := m.ErrorFallback.ToModelRoleConfig()
+ errorFallback = &c
+ }
+ var strongModel *ModelRoleConfig
+ if m.StrongModel != nil {
+ c := m.StrongModel.ToModelRoleConfig()
+ strongModel = &c
+ }
+
+ return ModelRoleConfig{
+ Role: m.Role,
+ BaseModelConfig: config.BaseModelConfig,
+ Temperature: m.Temperature,
+ TopP: m.TopP,
+ ReservedOutputTokens: m.ReservedOutputTokens,
+
+ LargeContextFallback: largeContextFallback,
+ LargeOutputFallback: largeOutputFallback,
+ ErrorFallback: errorFallback,
+ StrongModel: strongModel,
+ }
}
func (m ModelRoleConfig) GetReservedOutputTokens() int {
@@ -167,6 +271,20 @@ func (p PlannerRoleConfig) Value() (driver.Value, error) {
return json.Marshal(p)
}
+type ModelPackSchema struct {
+ Name string `json:"name"`
+ Description string `json:"description"`
+ Planner PlannerRoleConfig `json:"planner"`
+ Coder *ModelRoleConfig `json:"coder"`
+ PlanSummary ModelRoleConfig `json:"planSummary"`
+ Builder ModelRoleConfig `json:"builder"`
+ WholeFileBuilder *ModelRoleConfig `json:"wholeFileBuilder"` // optional, defaults to builder model — access via GetWholeFileBuilder()
+ Namer ModelRoleConfig `json:"namer"`
+ CommitMsg ModelRoleConfig `json:"commitMsg"`
+ ExecStatus ModelRoleConfig `json:"execStatus"`
+ Architect *ModelRoleConfig `json:"contextLoader"`
+}
+
type ModelPack struct {
Id string `json:"id"`
Name string `json:"name"`
diff --git a/app/shared/ai_models_providers.go b/app/shared/ai_models_providers.go
index d2d1568a..9a21f3b7 100644
--- a/app/shared/ai_models_providers.go
+++ b/app/shared/ai_models_providers.go
@@ -26,9 +26,20 @@ const (
ModelProviderDeepSeek ModelProvider = "deepseek"
ModelProviderPerplexity ModelProvider = "perplexity"
+ ModelProviderAmazonBedrock ModelProvider = "aws-bedrock"
+
+ ModelProviderOllama ModelProvider = "ollama"
+
ModelProviderCustom ModelProvider = "custom"
)
+var ModelProviderToLiteLLMId = map[ModelProvider]string{
+ ModelProviderGoogleAIStudio: "google_ai_studio",
+ ModelProviderGoogleVertex: "vertex_ai",
+ ModelProviderAzureOpenAI: "azure",
+ ModelProviderAmazonBedrock: "bedrock",
+}
+
var AllModelProviders = []string{
string(ModelProviderOpenAI),
string(ModelProviderOpenRouter),
@@ -38,69 +49,143 @@ var AllModelProviders = []string{
string(ModelProviderAzureOpenAI),
string(ModelProviderDeepSeek),
string(ModelProviderPerplexity),
+ string(ModelProviderAmazonBedrock),
string(ModelProviderCustom),
}
-var BaseUrlByProvider = map[ModelProvider]string{
- ModelProviderOpenAI: OpenAIV1BaseUrl,
- ModelProviderOpenRouter: OpenRouterBaseUrl,
-
- // apart from openai and openrouter, the rest are supported by liteLLM
- ModelProviderAnthropic: LiteLLMBaseUrl,
- ModelProviderGoogleAIStudio: LiteLLMBaseUrl,
- ModelProviderGoogleVertex: LiteLLMBaseUrl,
- ModelProviderAzureOpenAI: LiteLLMBaseUrl,
- ModelProviderDeepSeek: LiteLLMBaseUrl,
- ModelProviderPerplexity: LiteLLMBaseUrl,
-}
-
-var ApiKeyByProvider = map[ModelProvider]string{
- ModelProviderOpenAI: OpenAIEnvVar,
- ModelProviderOpenRouter: OpenRouterApiKeyEnvVar,
-
- ModelProviderAnthropic: AnthropicApiKeyEnvVar,
- ModelProviderGoogleAIStudio: GoogleAIStudioApiKeyEnvVar,
- ModelProviderGoogleVertex: GoogleVertexApiKeyEnvVar,
- ModelProviderAzureOpenAI: AzureOpenAIEnvVar,
- ModelProviderDeepSeek: DeepSeekApiKeyEnvVar,
- ModelProviderPerplexity: PerplexityApiKeyEnvVar,
-}
+// var BaseUrlByProvider = map[ModelProvider]string{
+// ModelProviderOpenAI: OpenAIV1BaseUrl,
+// ModelProviderOpenRouter: OpenRouterBaseUrl,
+
+// // apart from openai and openrouter, the rest are supported by liteLLM
+// ModelProviderAnthropic: LiteLLMBaseUrl,
+// ModelProviderGoogleAIStudio: LiteLLMBaseUrl,
+// ModelProviderGoogleVertex: LiteLLMBaseUrl,
+// ModelProviderAzureOpenAI: LiteLLMBaseUrl,
+// ModelProviderDeepSeek: LiteLLMBaseUrl,
+// ModelProviderPerplexity: LiteLLMBaseUrl,
+// ModelProviderAmazonBedrock: LiteLLMBaseUrl,
+// }
+
+// var ApiKeyByProvider = map[ModelProvider]string{
+// ModelProviderOpenAI: OpenAIEnvVar,
+// ModelProviderOpenRouter: OpenRouterApiKeyEnvVar,
+
+// ModelProviderAnthropic: AnthropicApiKeyEnvVar,
+// ModelProviderGoogleAIStudio: GoogleAIStudioApiKeyEnvVar,
+// ModelProviderGoogleVertex: GoogleVertexApiKeyEnvVar,
+// ModelProviderAzureOpenAI: AzureOpenAIEnvVar,
+// ModelProviderDeepSeek: DeepSeekApiKeyEnvVar,
+// ModelProviderPerplexity: PerplexityApiKeyEnvVar,
+// }
-// these types can come in via JSON config which uses JSON schema with discriminated unions — thus the need for the `omitempty` and pointers on all optional fields
type ModelProviderExtraAuthVars struct {
Var string `json:"var"`
- IsFilePath *bool `json:"isFilePath,omitempty"`
- Required *bool `json:"required,omitempty"`
+ IsFilePath bool `json:"isFilePath,omitempty"`
+ Required bool `json:"required,omitempty"`
+ Default string `json:"default,omitempty"`
}
-type ModelProviderConfig struct {
- SchemaVersion string `json:"schemaVersion"`
+type ModelProviderConfigSchema struct {
Provider ModelProvider `json:"provider"`
CustomProvider *string `json:"customProvider,omitempty"`
BaseUrl string `json:"baseUrl"`
+ // for AWS Bedrock models
+ HasAWSAuth bool `json:"hasAWSAuth,omitempty"`
+
// for local models that don't require auth (ollama, etc.)
- SkipAuth *bool `json:"skipAuth,omitempty"`
+ SkipAuth bool `json:"skipAuth,omitempty"`
- ApiKeyEnvVar *string `json:"apiKeyEnvVar,omitempty"`
- ExtraAuthVars *[]ModelProviderExtraAuthVars `json:"extraAuthVars,omitempty"`
+ ApiKeyEnvVar string `json:"apiKeyEnvVar,omitempty"`
+ ExtraAuthVars []ModelProviderExtraAuthVars `json:"extraAuthVars,omitempty"`
}
-var BuiltInModelProviderConfigs = map[ModelProvider]ModelProviderConfig{
- ModelProviderOpenAI: {
- Provider: ModelProviderOpenAI,
- BaseUrl: OpenAIV1BaseUrl,
+const DefaultAzureApiVersion = "2024-10-21"
+const AnthropicMaxReasoningBudget = 32000
- SkipAuth: &[]bool{false}[0],
- ApiKeyEnvVar: &[]string{OpenAIEnvVar}[0],
- ExtraAuthVars: &[]ModelProviderExtraAuthVars{},
+var BuiltInModelProviderConfigs = map[ModelProvider]ModelProviderConfigSchema{
+ ModelProviderOpenAI: {
+ Provider: ModelProviderOpenAI,
+ BaseUrl: OpenAIV1BaseUrl,
+ ApiKeyEnvVar: OpenAIEnvVar,
},
ModelProviderOpenRouter: {
- Provider: ModelProviderOpenRouter,
- BaseUrl: OpenRouterBaseUrl,
-
- SkipAuth: &[]bool{true}[0],
- ApiKeyEnvVar: &[]string{OpenRouterApiKeyEnvVar}[0],
- ExtraAuthVars: &[]ModelProviderExtraAuthVars{},
+ Provider: ModelProviderOpenRouter,
+ BaseUrl: OpenRouterBaseUrl,
+ ApiKeyEnvVar: OpenRouterApiKeyEnvVar,
+ },
+ ModelProviderAnthropic: {
+ Provider: ModelProviderAnthropic,
+ BaseUrl: LiteLLMBaseUrl,
+ ApiKeyEnvVar: AnthropicApiKeyEnvVar,
+ },
+ ModelProviderGoogleAIStudio: {
+ Provider: ModelProviderGoogleAIStudio,
+ BaseUrl: LiteLLMBaseUrl,
+ ApiKeyEnvVar: GoogleAIStudioApiKeyEnvVar,
+ },
+ ModelProviderGoogleVertex: {
+ Provider: ModelProviderGoogleVertex,
+ BaseUrl: LiteLLMBaseUrl,
+ ExtraAuthVars: []ModelProviderExtraAuthVars{
+ {
+ Var: "GOOGLE_APPLICATION_CREDENTIALS",
+ IsFilePath: true,
+ Required: true,
+ },
+ {
+ Var: "VERTEXAI_PROJECT",
+ Required: true,
+ },
+ {
+ Var: "VERTEXAI_LOCATION",
+ Required: true,
+ },
+ },
+ },
+ ModelProviderAzureOpenAI: {
+ Provider: ModelProviderAzureOpenAI,
+ BaseUrl: LiteLLMBaseUrl,
+ ExtraAuthVars: []ModelProviderExtraAuthVars{
+ {
+ Var: "AZURE_OPENAI_API_KEY",
+ Required: true,
+ },
+ {
+ Var: "AZURE_API_BASE",
+ Required: true,
+ },
+ {
+ Var: "AZURE_API_VERSION",
+ Required: false,
+ Default: DefaultAzureApiVersion,
+ },
+ },
+ },
+ ModelProviderDeepSeek: {
+ Provider: ModelProviderDeepSeek,
+ BaseUrl: LiteLLMBaseUrl,
+ ApiKeyEnvVar: DeepSeekApiKeyEnvVar,
+ },
+ ModelProviderPerplexity: {
+ Provider: ModelProviderPerplexity,
+ BaseUrl: LiteLLMBaseUrl,
+ ApiKeyEnvVar: PerplexityApiKeyEnvVar,
+ },
+ ModelProviderAmazonBedrock: {
+ Provider: ModelProviderAmazonBedrock,
+ BaseUrl: LiteLLMBaseUrl,
+ HasAWSAuth: true,
+ ExtraAuthVars: []ModelProviderExtraAuthVars{
+ {Var: "AWS_ACCESS_KEY_ID", Required: false},
+ {Var: "AWS_SECRET_ACCESS_KEY", Required: false},
+ {Var: "AWS_REGION", Required: false},
+ },
+ },
+ ModelProviderOllama: {
+ Provider: ModelProviderOllama,
+ BaseUrl: LiteLLMBaseUrl,
+ SkipAuth: true,
},
}
diff --git a/app/shared/json-schemas/available-model-packs.schema.json b/app/shared/json-schemas/available-model-packs.schema.json
new file mode 100644
index 00000000..e85edf24
--- /dev/null
+++ b/app/shared/json-schemas/available-model-packs.schema.json
@@ -0,0 +1,18 @@
+{
+ "$schema": "https://json-schema.org/draft/2020-12/schema",
+ "$id": "https://plandex.ai/schemas/available-model-packs.schema.json",
+ "title": "Available Model Packs",
+ "description": "A list of available model packs",
+ "type": "object",
+ "properties": {
+ "modelPacks": {
+ "type": "array",
+ "items": {
+ "$ref": "./model-pack-config.schema.json"
+ }
+ }
+ },
+ "required": [
+ "modelPacks"
+ ]
+}
\ No newline at end of file
diff --git a/app/shared/json-schemas/available-models.schema.json b/app/shared/json-schemas/available-models.schema.json
new file mode 100644
index 00000000..4c80d61e
--- /dev/null
+++ b/app/shared/json-schemas/available-models.schema.json
@@ -0,0 +1,18 @@
+{
+ "$schema": "https://json-schema.org/draft/2020-12/schema",
+ "$id": "https://plandex.ai/schemas/available-models.schema.json",
+ "title": "Available Models",
+ "description": "A list of available models",
+ "type": "object",
+ "properties": {
+ "models": {
+ "type": "array",
+ "items": {
+ "$ref": "./model-config.schema.json"
+ }
+ }
+ },
+ "required": [
+ "models"
+ ]
+}
\ No newline at end of file
diff --git a/app/shared/json-schemas/model-config.schema.json b/app/shared/json-schemas/model-config.schema.json
index 1a779c99..e8656972 100644
--- a/app/shared/json-schemas/model-config.schema.json
+++ b/app/shared/json-schemas/model-config.schema.json
@@ -5,10 +5,6 @@
"description": "Config for a model",
"type": "object",
"properties": {
- "schemaVersion": {
- "type": "string",
- "const": "2025-05-01"
- },
"modelTag": {
"type": "string"
},
@@ -70,13 +66,49 @@
"type": "array",
"items": {
"type": "object",
- "$ref": "./model-provider-config.schema.json",
- "minItems": 1
- }
+ "properties": {
+ "provider": {
+ "$ref": "./definitions/model-providers.schema.json"
+ },
+ "customProvider": {
+ "type": "string"
+ },
+ "modelName": {
+ "type": "string"
+ }
+ },
+ "required": [
+ "provider",
+ "modelName"
+ ],
+ "allOf": [
+ {
+ "if": {
+ "properties": {
+ "provider": {
+ "const": "custom"
+ }
+ }
+ },
+ "then": {
+ "required": [
+ "customProvider"
+ ]
+ },
+ "else": {
+ "not": {
+ "required": [
+ "customProvider"
+ ]
+ }
+ }
+ }
+ ]
+ },
+ "minItems": 1
}
},
"required": [
- "schemaVersion",
"modelTag",
"description",
"maxTokens",
diff --git a/app/shared/json-schemas/model-pack-config.schema.json b/app/shared/json-schemas/model-pack-config.schema.json
index 643e873c..1053b130 100644
--- a/app/shared/json-schemas/model-pack-config.schema.json
+++ b/app/shared/json-schemas/model-pack-config.schema.json
@@ -5,9 +5,68 @@
"description": "Config for a model pack",
"type": "object",
"properties": {
- "schemaVersion": {
+ "name": {
"type": "string",
- "const": "2025-05-01"
+ "description": "The name of the model pack"
+ },
+ "description": {
+ "type": "string",
+ "description": "The description of the model pack"
+ },
+ "planner": {
+ "type": "object",
+ "description": "The planner model config",
+ "$ref": "./model-role-config.schema.json"
+ },
+ "coder": {
+ "type": "object",
+ "description": "The coder model config",
+ "$ref": "./model-role-config.schema.json"
+ },
+ "architect": {
+ "type": "object",
+ "description": "The architect model config",
+ "$ref": "./model-role-config.schema.json"
+ },
+ "planSummary": {
+ "type": "object",
+ "description": "The plan summary model config",
+ "$ref": "./model-role-config.schema.json"
+ },
+ "builder": {
+ "type": "object",
+ "description": "The builder model config",
+ "$ref": "./model-role-config.schema.json"
+ },
+ "wholeFileBuilder": {
+ "type": "object",
+ "description": "The whole file builder model config",
+ "$ref": "./model-role-config.schema.json"
+ },
+ "namer": {
+ "type": "object",
+ "description": "The namer model config",
+ "$ref": "./model-role-config.schema.json"
+ },
+ "commitMsg": {
+ "type": "object",
+ "description": "The commit message model config",
+ "$ref": "./model-role-config.schema.json"
+ },
+ "execStatus": {
+ "type": "object",
+ "description": "The execution status model config",
+ "$ref": "./model-role-config.schema.json"
}
- }
+ },
+ "required": [
+ "name",
+ "description",
+ "planner",
+ "planSummary",
+ "builder",
+ "namer",
+ "commitMsg",
+ "execStatus"
+ ]
}
\ No newline at end of file
diff --git a/app/shared/json-schemas/model-provider-config.schema.json b/app/shared/json-schemas/model-provider-config.schema.json
index 7957cb3a..0cb8c4f7 100644
--- a/app/shared/json-schemas/model-provider-config.schema.json
+++ b/app/shared/json-schemas/model-provider-config.schema.json
@@ -5,10 +5,6 @@
"description": "Config for a model provider",
"type": "object",
"properties": {
- "schemaVersion": {
- "type": "string",
- "const": "2025-05-01"
- },
"provider": {
"$ref": "./definitions/model-providers.schema.json"
},
@@ -48,7 +44,6 @@
}
},
"required": [
- "schemaVersion",
"provider",
"baseUrl"
],
@@ -75,52 +70,37 @@
}
},
{
- "if": {
- "properties": {
- "skipAuth": {
- "const": false
- }
- }
- },
- "then": {
- "anyOf": [
- {
- "required": [
- "apiKeyEnvVar"
- ]
- },
- {
- "required": [
- "extraAuthVars"
- ]
- }
- ]
- }
- },
- {
- "if": {
- "properties": {
- "skipAuth": {
- "const": true
- }
- }
- },
- "then": {
- "not": {
- "anyOf": [
- {
- "required": [
- "apiKeyEnvVar"
- ]
- },
- {
- "required": [
- "extraAuthVars"
- ]
+ "anyOf": [
+ {
+ "properties": {
+ "provider": {
+ "not": {
+ "const": "custom"
+ }
}
+ }
+ },
+ {
+ "required": [
+ "apiKeyEnvVar"
]
+ },
+ {
+ "required": [
+ "extraAuthVars"
+ ]
+ },
+ {
+ "required": [
+ "skipAuth"
+ ],
+ "properties": {
+ "skipAuth": {
+ "const": true
+ }
+ }
}
- }
+ ]
}
]
}
\ No newline at end of file
diff --git a/app/shared/json-schemas/model-role-config.schema.json b/app/shared/json-schemas/model-role-config.schema.json
index 95054ba3..e57fd7ff 100644
--- a/app/shared/json-schemas/model-role-config.schema.json
+++ b/app/shared/json-schemas/model-role-config.schema.json
@@ -5,13 +5,12 @@
"description": "Config for a defined model role within a model pack",
"type": "object",
"properties": {
- "schemaVersion": {
- "type": "string",
- "const": "2025-05-01"
- },
"role": {
"$ref": "./definitions/model-roles.schema.json"
},
+ "modelTag": {
+ "type": "string"
+ },
"temperature": {
"type": "number"
},
@@ -20,10 +19,25 @@
},
"reservedOutputTokens": {
"type": "number"
+ },
+ "maxConvoTokens": {
+ "type": "number"
+ },
+ "largeContextFallback": {
+ "$ref": "./definitions/model-role-config.schema.json"
+ },
+ "largeOutputFallback": {
+ "$ref": "./definitions/model-role-config.schema.json"
+ },
+ "errorFallback": {
+ "$ref": "./definitions/model-role-config.schema.json"
+ },
+ "strongModel": {
+ "$ref": "./definitions/model-role-config.schema.json"
}
},
"required": [
- "schemaVersion",
- "role"
+ "role",
+ "modelTag"
]
}
\ No newline at end of file
diff --git a/app/shared/json-schemas/plan-config.schema.json b/app/shared/json-schemas/plan-config.schema.json
index 6b3a75fe..9a96f0a2 100644
--- a/app/shared/json-schemas/plan-config.schema.json
+++ b/app/shared/json-schemas/plan-config.schema.json
@@ -5,10 +5,6 @@
"description": "Config for a plan",
"type": "object",
"properties": {
- "schemaVersion": {
- "type": "string",
- "const": "2025-05-01"
- },
"autoMode": {
"$ref": "./definitions/auto-modes.schema.json"
}
diff --git a/app/shared/json/available_model_packs.json b/app/shared/json/available_model_packs.json
new file mode 100644
index 00000000..e69de29b
diff --git a/app/shared/json/available_models.json b/app/shared/json/available_models.json
new file mode 100644
index 00000000..74b61c2a
--- /dev/null
+++ b/app/shared/json/available_models.json
@@ -0,0 +1,28 @@
+{
+ "$schema": "../json-schemas/available-models.schema.json",
+ "models": [
+ {
+ "description": "OpenAI o3-high",
+ "modelTag": "openai/o3-high",
+ "maxTokens": 200000,
+ "maxOutputTokens": 100000,
+ "reservedOutputTokens": 40000,
+ "apiKeyEnvVar": "OPENAI_API_KEY",
+ "modelCompatibility": {
+ "hasImageSupport": true
+ },
+ "preferredModelOutputFormat": "xml",
+ "systemPromptDisabled": true,
+ "roleParamsDisabled": true,
+ "reasoningEffortEnabled": true,
+ "reasoningEffort": "high",
+ "stopDisabled": true,
+ "providers": [
+ {
+ "provider": "openai",
+ "modelName": "o3"
+ }
+ ]
+ }
+ ]
+}
\ No newline at end of file
diff --git a/app/shared/utils.go b/app/shared/utils.go
index 252b43dd..6c57c8ef 100644
--- a/app/shared/utils.go
+++ b/app/shared/utils.go
@@ -4,6 +4,7 @@ import (
"bytes"
"crypto/rand"
"fmt"
+ "reflect"
"regexp"
"strings"
"time"
@@ -139,6 +140,19 @@ func ReplaceReverse(s, old, new string, n int) string {
return res
}
+func Merge[T any](base T, ov T) T {
+ rvBase := reflect.ValueOf(&base).Elem() // addressable
+ rvOv := reflect.ValueOf(ov)
+
+ for i := 0; i < rvBase.NumField(); i++ {
+ fOv := rvOv.Field(i)
+ if !fOv.IsZero() { // ← built-in zero test
+ rvBase.Field(i).Set(fOv)
+ }
+ }
+ return base
+}
+
func NormalizeEOL(data []byte) []byte {
if !looksTextish(data) {
return data