diff --git a/src/api/providers/__tests__/anthropic-vertex.spec.ts b/src/api/providers/__tests__/anthropic-vertex.spec.ts index 9d83f265c7..c057b5eb55 100644 --- a/src/api/providers/__tests__/anthropic-vertex.spec.ts +++ b/src/api/providers/__tests__/anthropic-vertex.spec.ts @@ -705,7 +705,7 @@ describe("VertexHandler", () => { expect(result.temperature).toBe(1.0) }) - it("does not honor custom maxTokens for non-thinking models", () => { + it("honors custom maxTokens for all models", () => { const handler = new AnthropicVertexHandler({ apiKey: "test-api-key", apiModelId: "claude-3-7-sonnet@20250219", @@ -714,7 +714,7 @@ describe("VertexHandler", () => { }) const result = handler.getModel() - expect(result.maxTokens).toBe(8192) + expect(result.maxTokens).toBe(8192) // Capped to model's actual maxTokens expect(result.reasoningBudget).toBeUndefined() expect(result.temperature).toBe(0) }) diff --git a/src/api/providers/__tests__/anthropic.spec.ts b/src/api/providers/__tests__/anthropic.spec.ts index b1d0a2f6b3..1ae5390461 100644 --- a/src/api/providers/__tests__/anthropic.spec.ts +++ b/src/api/providers/__tests__/anthropic.spec.ts @@ -251,7 +251,7 @@ describe("AnthropicHandler", () => { expect(result.temperature).toBe(1.0) }) - it("does not honor custom maxTokens for non-thinking models", () => { + it("honors custom maxTokens for all models", () => { const handler = new AnthropicHandler({ apiKey: "test-api-key", apiModelId: "claude-3-7-sonnet-20250219", @@ -260,7 +260,7 @@ describe("AnthropicHandler", () => { }) const result = handler.getModel() - expect(result.maxTokens).toBe(8192) + expect(result.maxTokens).toBe(8192) // Capped to model's actual maxTokens expect(result.reasoningBudget).toBeUndefined() expect(result.temperature).toBe(0) }) diff --git a/src/api/providers/__tests__/openai.spec.ts b/src/api/providers/__tests__/openai.spec.ts index b4b5f29204..4810575786 100644 --- a/src/api/providers/__tests__/openai.spec.ts +++ b/src/api/providers/__tests__/openai.spec.ts @@ -263,14 +263,14 @@ describe("OpenAiHandler", () => { expect(callArgs.max_completion_tokens).toBeUndefined() }) - it("should use user-configured modelMaxTokens instead of model default maxTokens", async () => { + it("should use user-configured modelMaxTokens but cap it to model's max capability", async () => { const optionsWithUserMaxTokens: ApiHandlerOptions = { ...mockOptions, includeMaxTokens: true, - modelMaxTokens: 32000, // User-configured value + modelMaxTokens: 32000, // User tries to set higher than model supports openAiCustomModelInfo: { contextWindow: 128_000, - maxTokens: 4096, // Model's default value (should not be used) + maxTokens: 4096, // Model's actual max capability supportsPromptCache: false, }, } @@ -279,10 +279,32 @@ describe("OpenAiHandler", () => { // Consume the stream to trigger the API call for await (const _chunk of stream) { } - // Assert the mockCreate was called with user-configured modelMaxTokens (32000), not model default maxTokens (4096) + // Assert the mockCreate was called with the model's max capability (4096), not the user's request (32000) expect(mockCreate).toHaveBeenCalled() const callArgs = mockCreate.mock.calls[0][0] - expect(callArgs.max_completion_tokens).toBe(32000) + expect(callArgs.max_completion_tokens).toBe(4096) + }) + + it("should use user-configured modelMaxTokens when it's less than model's max", async () => { + const optionsWithLowerUserMaxTokens: ApiHandlerOptions = { + ...mockOptions, + includeMaxTokens: true, + modelMaxTokens: 2000, // User sets lower than model's max + openAiCustomModelInfo: { + contextWindow: 128_000, + maxTokens: 4096, // Model's max capability + supportsPromptCache: false, + }, + } + const handlerWithLowerMaxTokens = new OpenAiHandler(optionsWithLowerUserMaxTokens) + const stream = handlerWithLowerMaxTokens.createMessage(systemPrompt, messages) + // Consume the stream to trigger the API call + for await (const _chunk of stream) { + } + // Assert the mockCreate was called with user's setting (2000) + expect(mockCreate).toHaveBeenCalled() + const callArgs = mockCreate.mock.calls[0][0] + expect(callArgs.max_completion_tokens).toBe(2000) }) it("should fallback to model default maxTokens when user modelMaxTokens is not set", async () => { diff --git a/src/api/providers/__tests__/openrouter.spec.ts b/src/api/providers/__tests__/openrouter.spec.ts index ea850c47be..8d3d05a8c5 100644 --- a/src/api/providers/__tests__/openrouter.spec.ts +++ b/src/api/providers/__tests__/openrouter.spec.ts @@ -89,7 +89,7 @@ describe("OpenRouterHandler", () => { expect(result.info.supportsPromptCache).toBe(true) }) - it("honors custom maxTokens for thinking models", async () => { + it("honors custom maxTokens for all models", async () => { const handler = new OpenRouterHandler({ openRouterApiKey: "test-key", openRouterModelId: "anthropic/claude-3.7-sonnet:thinking", @@ -98,12 +98,12 @@ describe("OpenRouterHandler", () => { }) const result = await handler.fetchModel() - expect(result.maxTokens).toBe(128000) // Use actual implementation value + expect(result.maxTokens).toBe(32_768) // Not capped since model supports 128000 expect(result.reasoningBudget).toBeUndefined() // Use actual implementation value expect(result.temperature).toBe(0) // Use actual implementation value }) - it("does not honor custom maxTokens for non-thinking models", async () => { + it("honors custom maxTokens for non-thinking models", async () => { const handler = new OpenRouterHandler({ ...mockOptions, modelMaxTokens: 32_768, @@ -111,7 +111,7 @@ describe("OpenRouterHandler", () => { }) const result = await handler.fetchModel() - expect(result.maxTokens).toBe(8192) + expect(result.maxTokens).toBe(8192) // Capped to model's actual maxTokens expect(result.reasoningBudget).toBeUndefined() expect(result.temperature).toBe(0) }) diff --git a/src/api/providers/chutes.ts b/src/api/providers/chutes.ts index 62121bd19d..db32610dce 100644 --- a/src/api/providers/chutes.ts +++ b/src/api/providers/chutes.ts @@ -3,6 +3,7 @@ import { Anthropic } from "@anthropic-ai/sdk" import OpenAI from "openai" import type { ApiHandlerOptions } from "../../shared/api" +import { getModelMaxOutputTokens } from "../../shared/api" import { XmlMatcher } from "../../utils/xml-matcher" import { convertToR1Format } from "../transform/r1-format" import { convertToOpenAiMessages } from "../transform/openai-format" @@ -27,15 +28,17 @@ export class ChutesHandler extends BaseOpenAiCompatibleProvider { systemPrompt: string, messages: Anthropic.Messages.MessageParam[], ): OpenAI.Chat.Completions.ChatCompletionCreateParamsStreaming { - const { - id: model, - info: { maxTokens: max_tokens }, - } = this.getModel() + const model = this.getModel() + const max_tokens = getModelMaxOutputTokens({ + modelId: model.id, + model: model.info, + settings: this.options as any, + }) - const temperature = this.options.modelTemperature ?? this.getModel().info.temperature + const temperature = this.options.modelTemperature ?? model.info.temperature return { - model, + model: model.id, max_tokens, temperature, messages: [{ role: "system", content: systemPrompt }, ...convertToOpenAiMessages(messages)], diff --git a/src/api/providers/gemini.ts b/src/api/providers/gemini.ts index 6765c8676d..6ab441735f 100644 --- a/src/api/providers/gemini.ts +++ b/src/api/providers/gemini.ts @@ -71,7 +71,7 @@ export class GeminiHandler extends BaseProvider implements SingleCompletionHandl systemInstruction, httpOptions: this.options.googleGeminiBaseUrl ? { baseUrl: this.options.googleGeminiBaseUrl } : undefined, thinkingConfig, - maxOutputTokens: this.options.modelMaxTokens ?? maxTokens ?? undefined, + maxOutputTokens: maxTokens ?? undefined, temperature: this.options.modelTemperature ?? 0, } diff --git a/src/api/providers/glama.ts b/src/api/providers/glama.ts index 774d615709..eda7acf41b 100644 --- a/src/api/providers/glama.ts +++ b/src/api/providers/glama.ts @@ -5,7 +5,7 @@ import OpenAI from "openai" import { glamaDefaultModelId, glamaDefaultModelInfo, GLAMA_DEFAULT_TEMPERATURE } from "@roo-code/types" import { Package } from "../../shared/package" -import { ApiHandlerOptions } from "../../shared/api" +import { ApiHandlerOptions, getModelMaxOutputTokens } from "../../shared/api" import { ApiStream } from "../transform/stream" import { convertToOpenAiMessages } from "../transform/openai-format" @@ -49,12 +49,14 @@ export class GlamaHandler extends RouterProvider implements SingleCompletionHand addCacheBreakpoints(systemPrompt, openAiMessages) } - // Required by Anthropic; other providers default to max tokens allowed. - let maxTokens: number | undefined - - if (modelId.startsWith("anthropic/")) { - maxTokens = info.maxTokens ?? undefined - } + // Use getModelMaxOutputTokens to respect user's custom max tokens setting + const maxTokens = modelId.startsWith("anthropic/") + ? getModelMaxOutputTokens({ + modelId, + model: info, + settings: this.options as any, + }) + : undefined const requestOptions: OpenAI.Chat.ChatCompletionCreateParams = { model: modelId, @@ -130,7 +132,11 @@ export class GlamaHandler extends RouterProvider implements SingleCompletionHand } if (modelId.startsWith("anthropic/")) { - requestOptions.max_tokens = info.maxTokens + requestOptions.max_tokens = getModelMaxOutputTokens({ + modelId, + model: info, + settings: this.options as any, + }) } const response = await this.client.chat.completions.create(requestOptions) diff --git a/src/api/providers/lite-llm.ts b/src/api/providers/lite-llm.ts index e8cd58b12c..cbdb3d86b6 100644 --- a/src/api/providers/lite-llm.ts +++ b/src/api/providers/lite-llm.ts @@ -5,7 +5,7 @@ import { litellmDefaultModelId, litellmDefaultModelInfo } from "@roo-code/types" import { calculateApiCostOpenAI } from "../../shared/cost" -import { ApiHandlerOptions } from "../../shared/api" +import { ApiHandlerOptions, getModelMaxOutputTokens } from "../../shared/api" import { ApiStream, ApiStreamUsageChunk } from "../transform/stream" import { convertToOpenAiMessages } from "../transform/openai-format" @@ -44,8 +44,12 @@ export class LiteLLMHandler extends RouterProvider implements SingleCompletionHa ...convertToOpenAiMessages(messages), ] - // Required by some providers; others default to max tokens allowed - let maxTokens: number | undefined = info.maxTokens ?? undefined + // Use getModelMaxOutputTokens to respect user's custom max tokens setting + const maxTokens = getModelMaxOutputTokens({ + modelId, + model: info, + settings: this.options as any, + }) const requestOptions: OpenAI.Chat.Completions.ChatCompletionCreateParamsStreaming = { model: modelId, @@ -119,7 +123,11 @@ export class LiteLLMHandler extends RouterProvider implements SingleCompletionHa requestOptions.temperature = this.options.modelTemperature ?? 0 } - requestOptions.max_tokens = info.maxTokens + requestOptions.max_tokens = getModelMaxOutputTokens({ + modelId, + model: info, + settings: this.options as any, + }) const response = await this.client.chat.completions.create(requestOptions) return response.choices[0]?.message.content || "" diff --git a/src/api/providers/mistral.ts b/src/api/providers/mistral.ts index 7d48b9ef01..62d4065ef9 100644 --- a/src/api/providers/mistral.ts +++ b/src/api/providers/mistral.ts @@ -3,7 +3,7 @@ import { Mistral } from "@mistralai/mistralai" import { type MistralModelId, mistralDefaultModelId, mistralModels, MISTRAL_DEFAULT_TEMPERATURE } from "@roo-code/types" -import { ApiHandlerOptions } from "../../shared/api" +import { ApiHandlerOptions, getModelMaxOutputTokens } from "../../shared/api" import { convertToMistralMessages } from "../transform/mistral-format" import { ApiStream } from "../transform/stream" @@ -78,7 +78,13 @@ export class MistralHandler extends BaseProvider implements SingleCompletionHand const info = mistralModels[id as MistralModelId] ?? mistralModels[mistralDefaultModelId] // @TODO: Move this to the `getModelParams` function. - const maxTokens = this.options.includeMaxTokens ? info.maxTokens : undefined + const maxTokens = this.options.includeMaxTokens + ? getModelMaxOutputTokens({ + modelId: id, + model: info, + settings: { ...this.options, apiProvider: "mistral" } as any, + }) + : undefined const temperature = this.options.modelTemperature ?? MISTRAL_DEFAULT_TEMPERATURE return { id, info, maxTokens, temperature } diff --git a/src/api/providers/openai.ts b/src/api/providers/openai.ts index f5e4e4c985..3be3f679f9 100644 --- a/src/api/providers/openai.ts +++ b/src/api/providers/openai.ts @@ -10,7 +10,7 @@ import { OPENAI_AZURE_AI_INFERENCE_PATH, } from "@roo-code/types" -import type { ApiHandlerOptions } from "../../shared/api" +import { ApiHandlerOptions, getModelMaxOutputTokens } from "../../shared/api" import { XmlMatcher } from "../../utils/xml-matcher" @@ -403,9 +403,14 @@ export class OpenAiHandler extends BaseProvider implements SingleCompletionHandl ): void { // Only add max_completion_tokens if includeMaxTokens is true if (this.options.includeMaxTokens === true) { - // Use user-configured modelMaxTokens if available, otherwise fall back to model's default maxTokens + // Use getModelMaxOutputTokens to properly handle user settings and model limits // Using max_completion_tokens as max_tokens is deprecated - requestOptions.max_completion_tokens = this.options.modelMaxTokens || modelInfo.maxTokens + const modelId = this.options.openAiModelId ?? "" + requestOptions.max_completion_tokens = getModelMaxOutputTokens({ + modelId, + model: modelInfo, + settings: { ...this.options, apiProvider: "openai" } as any, + }) } } } diff --git a/src/api/providers/unbound.ts b/src/api/providers/unbound.ts index bc85dfd499..ee2e50a529 100644 --- a/src/api/providers/unbound.ts +++ b/src/api/providers/unbound.ts @@ -4,6 +4,7 @@ import OpenAI from "openai" import { unboundDefaultModelId, unboundDefaultModelInfo } from "@roo-code/types" import type { ApiHandlerOptions } from "../../shared/api" +import { getModelMaxOutputTokens } from "../../shared/api" import { ApiStream, ApiStreamUsageChunk } from "../transform/stream" import { convertToOpenAiMessages } from "../transform/openai-format" @@ -76,12 +77,14 @@ export class UnboundHandler extends RouterProvider implements SingleCompletionHa addVertexCacheBreakpoints(messages) } - // Required by Anthropic; other providers default to max tokens allowed. - let maxTokens: number | undefined - - if (modelId.startsWith("anthropic/")) { - maxTokens = info.maxTokens ?? undefined - } + // Use getModelMaxOutputTokens to respect user's custom max tokens setting + const maxTokens = modelId.startsWith("anthropic/") + ? getModelMaxOutputTokens({ + modelId, + model: info, + settings: this.options as any, + }) + : undefined const requestOptions: UnboundChatCompletionCreateParamsStreaming = { model: modelId.split("/")[1], @@ -149,7 +152,11 @@ export class UnboundHandler extends RouterProvider implements SingleCompletionHa } if (modelId.startsWith("anthropic/")) { - requestOptions.max_tokens = info.maxTokens + requestOptions.max_tokens = getModelMaxOutputTokens({ + modelId, + model: info, + settings: this.options as any, + }) } const response = await this.client.chat.completions.create(requestOptions, { headers: DEFAULT_HEADERS }) diff --git a/src/api/providers/xai.ts b/src/api/providers/xai.ts index 596c9e89b8..c1140804a2 100644 --- a/src/api/providers/xai.ts +++ b/src/api/providers/xai.ts @@ -3,7 +3,7 @@ import OpenAI from "openai" import { type XAIModelId, xaiDefaultModelId, xaiModels } from "@roo-code/types" -import type { ApiHandlerOptions } from "../../shared/api" +import { ApiHandlerOptions, getModelMaxOutputTokens } from "../../shared/api" import { ApiStream } from "../transform/stream" import { convertToOpenAiMessages } from "../transform/openai-format" @@ -50,7 +50,11 @@ export class XAIHandler extends BaseProvider implements SingleCompletionHandler // Use the OpenAI-compatible API. const stream = await this.client.chat.completions.create({ model: modelId, - max_tokens: modelInfo.maxTokens, + max_tokens: getModelMaxOutputTokens({ + modelId, + model: modelInfo, + settings: { ...this.options, apiProvider: "xai" } as any, + }), temperature: this.options.modelTemperature ?? XAI_DEFAULT_TEMPERATURE, messages: [{ role: "system", content: systemPrompt }, ...convertToOpenAiMessages(messages)], stream: true, @@ -78,12 +82,15 @@ export class XAIHandler extends BaseProvider implements SingleCompletionHandler if (chunk.usage) { // Extract detailed token information if available // First check for prompt_tokens_details structure (real API response) - const promptDetails = "prompt_tokens_details" in chunk.usage ? chunk.usage.prompt_tokens_details : null; - const cachedTokens = promptDetails && "cached_tokens" in promptDetails ? promptDetails.cached_tokens : 0; + const promptDetails = "prompt_tokens_details" in chunk.usage ? chunk.usage.prompt_tokens_details : null + const cachedTokens = promptDetails && "cached_tokens" in promptDetails ? promptDetails.cached_tokens : 0 // Fall back to direct fields in usage (used in test mocks) - const readTokens = cachedTokens || ("cache_read_input_tokens" in chunk.usage ? (chunk.usage as any).cache_read_input_tokens : 0); - const writeTokens = "cache_creation_input_tokens" in chunk.usage ? (chunk.usage as any).cache_creation_input_tokens : 0; + const readTokens = + cachedTokens || + ("cache_read_input_tokens" in chunk.usage ? (chunk.usage as any).cache_read_input_tokens : 0) + const writeTokens = + "cache_creation_input_tokens" in chunk.usage ? (chunk.usage as any).cache_creation_input_tokens : 0 yield { type: "usage", diff --git a/src/api/transform/__tests__/model-params.spec.ts b/src/api/transform/__tests__/model-params.spec.ts index c29d17559e..9fcba2959d 100644 --- a/src/api/transform/__tests__/model-params.spec.ts +++ b/src/api/transform/__tests__/model-params.spec.ts @@ -665,7 +665,7 @@ describe("getModelParams", () => { model, }) - expect(result.maxTokens).toBe(20000) + expect(result.maxTokens).toBe(16000) // Capped to model's actual maxTokens expect(result.reasoningBudget).toBe(10000) expect(result.temperature).toBe(1.0) // Overridden for reasoning budget models expect(result.reasoningEffort).toBeUndefined() // Budget takes precedence diff --git a/src/shared/__tests__/api.spec.ts b/src/shared/__tests__/api.spec.ts index 08d4bdf3bb..f75568b8a0 100644 --- a/src/shared/__tests__/api.spec.ts +++ b/src/shared/__tests__/api.spec.ts @@ -73,7 +73,8 @@ describe("getModelMaxOutputTokens", () => { settings, }) - expect(result).toBe(32000) + // Should cap to model's maxTokens (8192) even though user set 32000 + expect(result).toBe(8192) }) test("should return default of 8192 when maxTokens is undefined", () => { @@ -154,6 +155,93 @@ describe("getModelMaxOutputTokens", () => { expect(getModelMaxOutputTokens({ modelId: "test", model, settings })).toBe(16_384) }) + + test("should use user-configured modelMaxTokens for non-reasoning models", () => { + const settings: ProviderSettings = { + modelMaxTokens: 16000, + } + + const result = getModelMaxOutputTokens({ + modelId: "gpt-4", + model: mockModel, + settings, + }) + + // Should cap to model's maxTokens (8192) even though user set 16000 + expect(result).toBe(8192) + }) + + test("should ignore modelMaxTokens when it's 0 or negative", () => { + const settings: ProviderSettings = { + modelMaxTokens: 0, + } + + const result = getModelMaxOutputTokens({ + modelId: "claude-3-5-sonnet", + model: mockModel, + settings, + }) + + // Should fall back to model's maxTokens + expect(result).toBe(8192) + }) + + test("should prioritize user-configured modelMaxTokens over model's default", () => { + const modelWithHighMaxTokens: ModelInfo = { + maxTokens: 64000, + contextWindow: 200000, + supportsPromptCache: true, + } + + const settings: ProviderSettings = { + modelMaxTokens: 32000, + } + + const result = getModelMaxOutputTokens({ + modelId: "some-model", + model: modelWithHighMaxTokens, + settings, + }) + + expect(result).toBe(32000) + }) + + test("should use modelMaxTokens even for Anthropic models when configured", () => { + const settings: ProviderSettings = { + modelMaxTokens: 20000, + } + + const result = getModelMaxOutputTokens({ + modelId: "claude-3-5-sonnet", + model: mockModel, + settings, + format: "anthropic", + }) + + // Should cap to model's maxTokens (8192) even though user set 20000 + expect(result).toBe(8192) + }) + + test("should cap modelMaxTokens to model's actual max when user sets higher value", () => { + const model: ModelInfo = { + maxTokens: 16000, + contextWindow: 100000, + supportsPromptCache: true, + } + + const settings: ProviderSettings = { + modelMaxTokens: 32000, // Higher than model's max + } + + const result = getModelMaxOutputTokens({ + modelId: "some-model", + model, + settings, + }) + + // Should cap at model's max + expect(result).toBe(16000) + }) }) describe("shouldUseReasoningBudget", () => { diff --git a/src/shared/api.ts b/src/shared/api.ts index 8cbfc72133..f05e8c46a7 100644 --- a/src/shared/api.ts +++ b/src/shared/api.ts @@ -70,8 +70,15 @@ export const getModelMaxOutputTokens = ({ return settings.claudeCodeMaxOutputTokens || CLAUDE_CODE_DEFAULT_MAX_OUTPUT_TOKENS } + // Check for user-configured modelMaxTokens (takes precedence over all other logic) + if (settings?.modelMaxTokens && settings.modelMaxTokens > 0) { + const maxAllowed = model?.maxTokens || Number.MAX_SAFE_INTEGER + return Math.min(settings.modelMaxTokens, maxAllowed) + } + + // Existing reasoning budget logic if (shouldUseReasoningBudget({ model, settings })) { - return settings?.modelMaxTokens || DEFAULT_HYBRID_REASONING_MODEL_MAX_TOKENS + return DEFAULT_HYBRID_REASONING_MODEL_MAX_TOKENS } const isAnthropicContext = diff --git a/webview-ui/src/components/settings/ApiOptions.tsx b/webview-ui/src/components/settings/ApiOptions.tsx index 06994b16b9..788219e8b8 100644 --- a/webview-ui/src/components/settings/ApiOptions.tsx +++ b/webview-ui/src/components/settings/ApiOptions.tsx @@ -81,6 +81,7 @@ import { DiffSettingsControl } from "./DiffSettingsControl" import { TemperatureControl } from "./TemperatureControl" import { RateLimitSecondsControl } from "./RateLimitSecondsControl" import { ConsecutiveMistakeLimitControl } from "./ConsecutiveMistakeLimitControl" +import { MaxTokensSlider } from "./MaxTokensSlider" import { BedrockCustomArn } from "./providers/BedrockCustomArn" import { buildDocLink } from "@src/utils/docLinks" @@ -574,6 +575,13 @@ const ApiOptions = ({ onChange={handleInputChange("modelTemperature", noTransform)} maxValue={2} /> + {selectedProvider !== "openai" && ( + setApiConfigurationField("modelMaxTokens", value)} + modelInfo={selectedModelInfo} + /> + )} setApiConfigurationField("rateLimitSeconds", value)} diff --git a/webview-ui/src/components/settings/MaxTokensSlider.tsx b/webview-ui/src/components/settings/MaxTokensSlider.tsx new file mode 100644 index 0000000000..b34e8de82c --- /dev/null +++ b/webview-ui/src/components/settings/MaxTokensSlider.tsx @@ -0,0 +1,46 @@ +import React from "react" +import { useAppTranslation } from "@/i18n/TranslationContext" +import { ModelInfo } from "@roo-code/types" +import { Slider } from "@/components/ui" +import { DEFAULT_HYBRID_REASONING_MODEL_MAX_TOKENS } from "@roo/api" + +interface MaxTokensSliderProps { + value?: number + onChange: (value: number | undefined) => void + modelInfo?: ModelInfo + className?: string +} + +export const MaxTokensSlider: React.FC = ({ value, onChange, modelInfo, className }) => { + const { t } = useAppTranslation() + + // Use the same logic as the original ThinkingBudget component + const customMaxOutputTokens = value || DEFAULT_HYBRID_REASONING_MODEL_MAX_TOKENS + const maxValue = modelInfo?.maxTokens + ? Math.max(modelInfo.maxTokens, customMaxOutputTokens, DEFAULT_HYBRID_REASONING_MODEL_MAX_TOKENS) + : Math.max(customMaxOutputTokens, DEFAULT_HYBRID_REASONING_MODEL_MAX_TOKENS) + + return ( +
+
{t("settings:providers.maxOutputTokens.label")}
+
+ onChange(value)} + /> +
{customMaxOutputTokens}
+
+
+ {t("settings:providers.maxOutputTokens.description")} +
+ {modelInfo && ( +
+ {t("settings:providers.maxOutputTokens.modelSupports", { max: modelInfo.maxTokens })} +
+ )} +
+ ) +} diff --git a/webview-ui/src/components/settings/ThinkingBudget.tsx b/webview-ui/src/components/settings/ThinkingBudget.tsx index 0adb62f2a0..d37e24db5e 100644 --- a/webview-ui/src/components/settings/ThinkingBudget.tsx +++ b/webview-ui/src/components/settings/ThinkingBudget.tsx @@ -59,38 +59,19 @@ export const ThinkingBudget = ({ apiConfiguration, setApiConfigurationField, mod )} {(isReasoningBudgetRequired || enableReasoningEffort) && ( - <> -
-
{t("settings:thinkingBudget.maxTokens")}
-
- setApiConfigurationField("modelMaxTokens", value)} - /> -
{customMaxOutputTokens}
-
-
-
-
{t("settings:thinkingBudget.maxThinkingTokens")}
-
- setApiConfigurationField("modelMaxThinkingTokens", value)} - /> -
{customMaxThinkingTokens}
-
+
+
{t("settings:thinkingBudget.maxThinkingTokens")}
+
+ setApiConfigurationField("modelMaxThinkingTokens", value)} + /> +
{customMaxThinkingTokens}
- +
)} ) : isReasoningEffortSupported ? ( diff --git a/webview-ui/src/components/settings/__tests__/ApiOptions.spec.tsx b/webview-ui/src/components/settings/__tests__/ApiOptions.spec.tsx index bf02840429..6beb192035 100644 --- a/webview-ui/src/components/settings/__tests__/ApiOptions.spec.tsx +++ b/webview-ui/src/components/settings/__tests__/ApiOptions.spec.tsx @@ -83,9 +83,21 @@ vi.mock("@/components/ui", () => ({ Popover: ({ children, _open, _onOpenChange }: any) =>
{children}
, PopoverContent: ({ children, _className }: any) =>
{children}
, PopoverTrigger: ({ children, _asChild }: any) =>
{children}
, - Slider: ({ value, onChange }: any) => ( + Slider: ({ value, onChange, onValueChange, min, max, step }: any) => (
- onChange(parseFloat(e.target.value))} /> + { + const val = parseFloat(e.target.value) + if (onChange) onChange(val) + if (onValueChange) onValueChange([val]) + }} + data-testid="slider-input" + />
), SearchableSelect: ({ value, onValueChange, options, placeholder, "data-testid": dataTestId }: any) => ( @@ -114,6 +126,19 @@ vi.mock("@/components/ui", () => ({ CollapsibleContent: ({ children, className }: any) => (
{children}
), + // Add Input component for other controls + Input: ({ id, type, value, onChange, min, max, className, ...props }: any) => ( + + ), })) vi.mock("../TemperatureControl", () => ({ diff --git a/webview-ui/src/components/settings/__tests__/MaxTokensSlider.spec.tsx b/webview-ui/src/components/settings/__tests__/MaxTokensSlider.spec.tsx new file mode 100644 index 0000000000..e9dbf6cd42 --- /dev/null +++ b/webview-ui/src/components/settings/__tests__/MaxTokensSlider.spec.tsx @@ -0,0 +1,107 @@ +import React from "react" +import { render, screen, fireEvent } from "@testing-library/react" +import { describe, it, expect, vi, beforeEach } from "vitest" +import { MaxTokensSlider } from "../MaxTokensSlider" +import { ModelInfo } from "@roo-code/types" + +// Mock the translation hook +vi.mock("@/i18n/TranslationContext", () => ({ + useAppTranslation: () => ({ + t: (key: string, params?: any) => { + if (key === "settings:providers.maxOutputTokens.label") return "Max Output Tokens" + if (key === "settings:providers.maxOutputTokens.description") return "Maximum number of tokens" + if (key === "settings:providers.maxOutputTokens.modelSupports") + return `Model supports up to ${params?.max} tokens` + return key + }, + }), +})) + +// Mock the Slider component +vi.mock("@/components/ui", () => ({ + Slider: ({ value, onValueChange, min, max, step }: any) => ( +
+ onValueChange([parseInt(e.target.value)])} + data-testid="slider-input" + /> +
+ ), +})) + +describe("MaxTokensSlider", () => { + const mockOnChange = vi.fn() + + beforeEach(() => { + mockOnChange.mockClear() + }) + + it("renders with default value", () => { + render() + + expect(screen.getByText("Max Output Tokens")).toBeInTheDocument() + expect(screen.getByText("Maximum number of tokens")).toBeInTheDocument() + // Default value is DEFAULT_HYBRID_REASONING_MODEL_MAX_TOKENS which is 16384 + expect(screen.getByText("16384")).toBeInTheDocument() + }) + + it("renders with custom value", () => { + render() + + expect(screen.getByText("16384")).toBeInTheDocument() + }) + + it("shows model support information when modelInfo is provided", () => { + const modelInfo: ModelInfo = { + maxTokens: 32768, + contextWindow: 100000, + supportsPromptCache: false, + } + + render() + + expect(screen.getByText("Model supports up to 32768 tokens")).toBeInTheDocument() + }) + + it("calls onChange when slider value changes", () => { + render() + + const slider = screen.getByTestId("slider-input") as HTMLInputElement + // Simulate changing the slider value + fireEvent.change(slider, { target: { value: "16384" } }) + + expect(mockOnChange).toHaveBeenCalledWith(16384) + }) + + it("calculates max value correctly based on model info", () => { + const modelInfo: ModelInfo = { + maxTokens: 65536, + contextWindow: 100000, + supportsPromptCache: false, + } + + render() + + const slider = screen.getByTestId("slider-input") as HTMLInputElement + expect(slider.max).toBe("65536") + }) + + it("uses default max value when model info is not provided", () => { + render() + + const slider = screen.getByTestId("slider-input") as HTMLInputElement + // When no model info, max is Math.max(value, DEFAULT_HYBRID_REASONING_MODEL_MAX_TOKENS) + expect(slider.max).toBe("16384") + }) + + it("applies custom className", () => { + const { container } = render() + + expect(container.firstChild).toHaveClass("custom-class") + }) +}) diff --git a/webview-ui/src/components/settings/__tests__/ThinkingBudget.spec.tsx b/webview-ui/src/components/settings/__tests__/ThinkingBudget.spec.tsx index 5ca51b4528..44dc291fd3 100644 --- a/webview-ui/src/components/settings/__tests__/ThinkingBudget.spec.tsx +++ b/webview-ui/src/components/settings/__tests__/ThinkingBudget.spec.tsx @@ -57,10 +57,10 @@ describe("ThinkingBudget", () => { expect(container.firstChild).toBeNull() }) - it("should render sliders when model supports thinking", () => { + it("should render slider when model supports thinking", () => { render() - expect(screen.getAllByTestId("slider")).toHaveLength(2) + expect(screen.getAllByTestId("slider")).toHaveLength(1) }) it("should update modelMaxThinkingTokens", () => { @@ -74,8 +74,8 @@ describe("ThinkingBudget", () => { />, ) - const sliders = screen.getAllByTestId("slider") - fireEvent.change(sliders[1], { target: { value: "5000" } }) + const slider = screen.getByTestId("slider") + fireEvent.change(slider, { target: { value: "5000" } }) expect(setApiConfigurationField).toHaveBeenCalledWith("modelMaxThinkingTokens", 5000) }) @@ -95,35 +95,39 @@ describe("ThinkingBudget", () => { expect(setApiConfigurationField).toHaveBeenCalledWith("modelMaxThinkingTokens", 8000) // 80% of 10000 }) + it("should allow max thinking tokens up to 80% of max output tokens", () => { + render() + + const slider = screen.getByTestId("slider") + // Max should be 80% of 5000 = 4000 + expect(slider.getAttribute("max")).toBe("4000") + }) + it("should use default thinking tokens if not provided", () => { render() // Default is 80% of max tokens, capped at 8192 - const sliders = screen.getAllByTestId("slider") - expect(sliders[1]).toHaveValue("8000") // 80% of 10000 + const slider = screen.getByTestId("slider") + expect(slider).toHaveValue("8000") // 80% of 10000 }) it("should use min thinking tokens of 1024", () => { - render() + render() - const sliders = screen.getAllByTestId("slider") - expect(sliders[1].getAttribute("min")).toBe("1024") + const slider = screen.getByTestId("slider") + expect(slider.getAttribute("min")).toBe("1024") }) - it("should update max tokens when slider changes", () => { - const setApiConfigurationField = vi.fn() - + it("should cap displayed value at 80% even if stored value is higher", () => { render( , ) - const sliders = screen.getAllByTestId("slider") - fireEvent.change(sliders[0], { target: { value: "12000" } }) - - expect(setApiConfigurationField).toHaveBeenCalledWith("modelMaxTokens", 12000) + const slider = screen.getByTestId("slider") + // Value should be capped at 4000 (80% of 5000) even though stored value is 8192 + expect(slider).toHaveValue("4000") }) }) diff --git a/webview-ui/src/i18n/locales/ca/settings.json b/webview-ui/src/i18n/locales/ca/settings.json index 207fc88fd9..bda230d575 100644 --- a/webview-ui/src/i18n/locales/ca/settings.json +++ b/webview-ui/src/i18n/locales/ca/settings.json @@ -386,6 +386,15 @@ "placeholder": "Per defecte: claude", "maxTokensLabel": "Tokens màxims de sortida", "maxTokensDescription": "Nombre màxim de tokens de sortida per a les respostes de Claude Code. El valor per defecte és 8000." + }, + "maxOutputTokens": { + "label": "Tokens màxims de sortida", + "description": "Nombre màxim de tokens a reservar per a la sortida del model. El valor predeterminat és 8192.", + "validation": { + "tooHigh": "El valor excedeix el màxim del model de {{max}} tokens", + "tooLow": "El valor ha de ser almenys {{min}} tokens" + }, + "modelSupports": "Aquest model admet fins a {{max}} tokens" } }, "browser": { @@ -654,7 +663,6 @@ } }, "thinkingBudget": { - "maxTokens": "Tokens màxims", "maxThinkingTokens": "Tokens de pensament màxims" }, "validation": { diff --git a/webview-ui/src/i18n/locales/de/settings.json b/webview-ui/src/i18n/locales/de/settings.json index 68e9f8f97d..3d08c1f99d 100644 --- a/webview-ui/src/i18n/locales/de/settings.json +++ b/webview-ui/src/i18n/locales/de/settings.json @@ -386,6 +386,15 @@ "placeholder": "Standard: claude", "maxTokensLabel": "Maximale Ausgabe-Tokens", "maxTokensDescription": "Maximale Anzahl an Ausgabe-Tokens für Claude Code-Antworten. Standard ist 8000." + }, + "maxOutputTokens": { + "label": "Maximale Ausgabe-Token", + "description": "Maximale Anzahl von Token, die für die Modellausgabe reserviert werden sollen. Standard ist 8192.", + "validation": { + "tooHigh": "Wert überschreitet das Maximum des Modells von {{max}} Token", + "tooLow": "Wert muss mindestens {{min}} Token betragen" + }, + "modelSupports": "Dieses Modell unterstützt bis zu {{max}} Token" } }, "browser": { @@ -654,7 +663,6 @@ } }, "thinkingBudget": { - "maxTokens": "Max Tokens", "maxThinkingTokens": "Max Thinking Tokens" }, "validation": { diff --git a/webview-ui/src/i18n/locales/en/settings.json b/webview-ui/src/i18n/locales/en/settings.json index fa1fbab13c..e4900fedf2 100644 --- a/webview-ui/src/i18n/locales/en/settings.json +++ b/webview-ui/src/i18n/locales/en/settings.json @@ -367,6 +367,15 @@ "label": "Rate limit", "description": "Minimum time between API requests." }, + "maxOutputTokens": { + "label": "Max Output Tokens", + "description": "Maximum number of tokens to reserve for model output. Default is 8192.", + "validation": { + "tooHigh": "Value exceeds model's maximum of {{max}} tokens", + "tooLow": "Value must be at least {{min}} tokens" + }, + "modelSupports": "This model supports up to {{max}} tokens" + }, "consecutiveMistakeLimit": { "label": "Error & Repetition Limit", "description": "Number of consecutive errors or repeated actions before showing 'Roo is having trouble' dialog", @@ -654,7 +663,6 @@ } }, "thinkingBudget": { - "maxTokens": "Max Tokens", "maxThinkingTokens": "Max Thinking Tokens" }, "validation": { diff --git a/webview-ui/src/i18n/locales/es/settings.json b/webview-ui/src/i18n/locales/es/settings.json index 80de7042b0..48d80fd510 100644 --- a/webview-ui/src/i18n/locales/es/settings.json +++ b/webview-ui/src/i18n/locales/es/settings.json @@ -367,6 +367,15 @@ "label": "Límite de tasa", "description": "Tiempo mínimo entre solicitudes de API." }, + "maxOutputTokens": { + "label": "Tokens máximos de salida", + "description": "Número máximo de tokens a reservar para la salida del modelo. El valor predeterminado es 8192.", + "validation": { + "tooHigh": "El valor excede el máximo del modelo de {{max}} tokens", + "tooLow": "El valor debe ser al menos {{min}} tokens" + }, + "modelSupports": "Este modelo admite hasta {{max}} tokens" + }, "consecutiveMistakeLimit": { "label": "Límite de errores y repeticiones", "description": "Número de errores consecutivos o acciones repetidas antes de mostrar el diálogo 'Roo está teniendo problemas'", @@ -654,7 +663,6 @@ } }, "thinkingBudget": { - "maxTokens": "Tokens máximos", "maxThinkingTokens": "Tokens máximos de pensamiento" }, "validation": { diff --git a/webview-ui/src/i18n/locales/fr/settings.json b/webview-ui/src/i18n/locales/fr/settings.json index e10b3ae35f..2c2346ed48 100644 --- a/webview-ui/src/i18n/locales/fr/settings.json +++ b/webview-ui/src/i18n/locales/fr/settings.json @@ -386,6 +386,15 @@ "placeholder": "Défaut : claude", "maxTokensLabel": "Jetons de sortie max", "maxTokensDescription": "Nombre maximum de jetons de sortie pour les réponses de Claude Code. La valeur par défaut est 8000." + }, + "maxOutputTokens": { + "label": "Tokens de sortie maximum", + "description": "Nombre maximum de tokens à réserver pour la sortie du modèle. La valeur par défaut est 8192.", + "validation": { + "tooHigh": "La valeur dépasse le maximum du modèle de {{max}} tokens", + "tooLow": "La valeur doit être d'au moins {{min}} tokens" + }, + "modelSupports": "Ce modèle prend en charge jusqu'à {{max}} tokens" } }, "browser": { @@ -654,7 +663,6 @@ } }, "thinkingBudget": { - "maxTokens": "Tokens maximum", "maxThinkingTokens": "Tokens de réflexion maximum" }, "validation": { diff --git a/webview-ui/src/i18n/locales/hi/settings.json b/webview-ui/src/i18n/locales/hi/settings.json index 5b64359f8f..5387175f5c 100644 --- a/webview-ui/src/i18n/locales/hi/settings.json +++ b/webview-ui/src/i18n/locales/hi/settings.json @@ -386,6 +386,15 @@ "placeholder": "डिफ़ॉल्ट: claude", "maxTokensLabel": "अधिकतम आउटपुट टोकन", "maxTokensDescription": "Claude Code प्रतिक्रियाओं के लिए आउटपुट टोकन की अधिकतम संख्या। डिफ़ॉल्ट 8000 है।" + }, + "maxOutputTokens": { + "label": "अधिकतम आउटपुट टोकन", + "description": "मॉडल आउटपुट के लिए आरक्षित करने के लिए टोकन की अधिकतम संख्या। डिफ़ॉल्ट 8192 है।", + "validation": { + "tooHigh": "मान मॉडल के अधिकतम {{max}} टोकन से अधिक है", + "tooLow": "मान कम से कम {{min}} टोकन होना चाहिए" + }, + "modelSupports": "यह मॉडल {{max}} टोकन तक का समर्थन करता है" } }, "browser": { @@ -654,7 +663,6 @@ } }, "thinkingBudget": { - "maxTokens": "अधिकतम tokens", "maxThinkingTokens": "अधिकतम thinking tokens" }, "validation": { diff --git a/webview-ui/src/i18n/locales/id/settings.json b/webview-ui/src/i18n/locales/id/settings.json index 24404dec3d..41707c3707 100644 --- a/webview-ui/src/i18n/locales/id/settings.json +++ b/webview-ui/src/i18n/locales/id/settings.json @@ -390,6 +390,15 @@ "placeholder": "Default: claude", "maxTokensLabel": "Token Output Maks", "maxTokensDescription": "Jumlah maksimum token output untuk respons Claude Code. Default adalah 8000." + }, + "maxOutputTokens": { + "label": "Token Output Maksimum", + "description": "Jumlah maksimum token yang akan dipesan untuk output model. Default adalah 8192.", + "validation": { + "tooHigh": "Nilai melebihi maksimum model {{max}} token", + "tooLow": "Nilai harus minimal {{min}} token" + }, + "modelSupports": "Model ini mendukung hingga {{max}} token" } }, "browser": { @@ -683,7 +692,6 @@ } }, "thinkingBudget": { - "maxTokens": "Token Maksimum", "maxThinkingTokens": "Token Thinking Maksimum" }, "validation": { diff --git a/webview-ui/src/i18n/locales/it/settings.json b/webview-ui/src/i18n/locales/it/settings.json index bacf7a40ef..72b5fb0cb4 100644 --- a/webview-ui/src/i18n/locales/it/settings.json +++ b/webview-ui/src/i18n/locales/it/settings.json @@ -386,6 +386,15 @@ "placeholder": "Predefinito: claude", "maxTokensLabel": "Token di output massimi", "maxTokensDescription": "Numero massimo di token di output per le risposte di Claude Code. Il valore predefinito è 8000." + }, + "maxOutputTokens": { + "label": "Token di output massimi", + "description": "Numero massimo di token da riservare per l'output del modello. Il valore predefinito è 8192.", + "validation": { + "tooHigh": "Il valore supera il massimo del modello di {{max}} token", + "tooLow": "Il valore deve essere almeno {{min}} token" + }, + "modelSupports": "Questo modello supporta fino a {{max}} token" } }, "browser": { @@ -654,7 +663,6 @@ } }, "thinkingBudget": { - "maxTokens": "Token massimi", "maxThinkingTokens": "Token massimi di pensiero" }, "validation": { diff --git a/webview-ui/src/i18n/locales/ja/settings.json b/webview-ui/src/i18n/locales/ja/settings.json index c588edead2..490c92c8c9 100644 --- a/webview-ui/src/i18n/locales/ja/settings.json +++ b/webview-ui/src/i18n/locales/ja/settings.json @@ -386,6 +386,15 @@ "placeholder": "デフォルト：claude", "maxTokensLabel": "最大出力トークン", "maxTokensDescription": "Claude Codeレスポンスの最大出力トークン数。デフォルトは8000です。" + }, + "maxOutputTokens": { + "label": "最大出力トークン", + "description": "モデル出力用に予約するトークンの最大数。デフォルトは8192です。", + "validation": { + "tooHigh": "値がモデルの最大値{{max}}トークンを超えています", + "tooLow": "値は最低{{min}}トークンである必要があります" + }, + "modelSupports": "このモデルは最大{{max}}トークンをサポートします" } }, "browser": { @@ -654,7 +663,6 @@ } }, "thinkingBudget": { - "maxTokens": "最大 tokens", "maxThinkingTokens": "最大思考 tokens" }, "validation": { diff --git a/webview-ui/src/i18n/locales/ko/settings.json b/webview-ui/src/i18n/locales/ko/settings.json index 84c329ffd7..024fe6b544 100644 --- a/webview-ui/src/i18n/locales/ko/settings.json +++ b/webview-ui/src/i18n/locales/ko/settings.json @@ -386,6 +386,15 @@ "placeholder": "기본값: claude", "maxTokensLabel": "최대 출력 토큰", "maxTokensDescription": "Claude Code 응답의 최대 출력 토큰 수. 기본값은 8000입니다." + }, + "maxOutputTokens": { + "label": "최대 출력 토큰", + "description": "모델 출력을 위해 예약할 최대 토큰 수입니다. 기본값은 8192입니다.", + "validation": { + "tooHigh": "값이 모델의 최대값 {{max}} 토큰을 초과합니다", + "tooLow": "값은 최소 {{min}} 토큰이어야 합니다" + }, + "modelSupports": "이 모델은 최대 {{max}} 토큰을 지원합니다" } }, "browser": { @@ -654,7 +663,6 @@ } }, "thinkingBudget": { - "maxTokens": "최대 tokens", "maxThinkingTokens": "최대 사고 tokens" }, "validation": { diff --git a/webview-ui/src/i18n/locales/nl/settings.json b/webview-ui/src/i18n/locales/nl/settings.json index 343008ae51..b9a239367d 100644 --- a/webview-ui/src/i18n/locales/nl/settings.json +++ b/webview-ui/src/i18n/locales/nl/settings.json @@ -386,6 +386,15 @@ "placeholder": "Standaard: claude", "maxTokensLabel": "Max Output Tokens", "maxTokensDescription": "Maximaal aantal output-tokens voor Claude Code-reacties. Standaard is 8000." + }, + "maxOutputTokens": { + "label": "Maximale uitvoertokens", + "description": "Maximum aantal tokens om te reserveren voor modeluitvoer. Standaard is 8192.", + "validation": { + "tooHigh": "Waarde overschrijdt het maximum van het model van {{max}} tokens", + "tooLow": "Waarde moet minimaal {{min}} tokens zijn" + }, + "modelSupports": "Dit model ondersteunt tot {{max}} tokens" } }, "browser": { @@ -654,7 +663,6 @@ } }, "thinkingBudget": { - "maxTokens": "Max tokens", "maxThinkingTokens": "Max denk-tokens" }, "validation": { diff --git a/webview-ui/src/i18n/locales/pl/settings.json b/webview-ui/src/i18n/locales/pl/settings.json index 2f300384e4..981fae0625 100644 --- a/webview-ui/src/i18n/locales/pl/settings.json +++ b/webview-ui/src/i18n/locales/pl/settings.json @@ -386,6 +386,15 @@ "placeholder": "Domyślnie: claude", "maxTokensLabel": "Maksymalna liczba tokenów wyjściowych", "maxTokensDescription": "Maksymalna liczba tokenów wyjściowych dla odpowiedzi Claude Code. Domyślnie 8000." + }, + "maxOutputTokens": { + "label": "Maksymalne tokeny wyjściowe", + "description": "Maksymalna liczba tokenów do zarezerwowania dla wyjścia modelu. Domyślnie 8192.", + "validation": { + "tooHigh": "Wartość przekracza maksimum modelu {{max}} tokenów", + "tooLow": "Wartość musi wynosić co najmniej {{min}} tokenów" + }, + "modelSupports": "Ten model obsługuje do {{max}} tokenów" } }, "browser": { @@ -654,7 +663,6 @@ } }, "thinkingBudget": { - "maxTokens": "Maksymalna liczba tokenów", "maxThinkingTokens": "Maksymalna liczba tokenów myślenia" }, "validation": { diff --git a/webview-ui/src/i18n/locales/pt-BR/settings.json b/webview-ui/src/i18n/locales/pt-BR/settings.json index 11e4d01aac..656f8ab1cb 100644 --- a/webview-ui/src/i18n/locales/pt-BR/settings.json +++ b/webview-ui/src/i18n/locales/pt-BR/settings.json @@ -386,6 +386,15 @@ "placeholder": "Padrão: claude", "maxTokensLabel": "Tokens de saída máximos", "maxTokensDescription": "Número máximo de tokens de saída para respostas do Claude Code. O padrão é 8000." + }, + "maxOutputTokens": { + "label": "Tokens máximos de saída", + "description": "Número máximo de tokens a reservar para a saída do modelo. O padrão é 8192.", + "validation": { + "tooHigh": "O valor excede o máximo do modelo de {{max}} tokens", + "tooLow": "O valor deve ser pelo menos {{min}} tokens" + }, + "modelSupports": "Este modelo suporta até {{max}} tokens" } }, "browser": { @@ -654,7 +663,6 @@ } }, "thinkingBudget": { - "maxTokens": "Tokens máximos", "maxThinkingTokens": "Tokens máximos de pensamento" }, "validation": { diff --git a/webview-ui/src/i18n/locales/ru/settings.json b/webview-ui/src/i18n/locales/ru/settings.json index 35eb9b1966..50b17808b9 100644 --- a/webview-ui/src/i18n/locales/ru/settings.json +++ b/webview-ui/src/i18n/locales/ru/settings.json @@ -386,6 +386,15 @@ "placeholder": "По умолчанию: claude", "maxTokensLabel": "Макс. выходных токенов", "maxTokensDescription": "Максимальное количество выходных токенов для ответов Claude Code. По умолчанию 8000." + }, + "maxOutputTokens": { + "label": "Максимальное количество выходных токенов", + "description": "Максимальное количество токенов для резервирования для вывода модели. По умолчанию 8192.", + "validation": { + "tooHigh": "Значение превышает максимум модели {{max}} токенов", + "tooLow": "Значение должно быть не менее {{min}} токенов" + }, + "modelSupports": "Эта модель поддерживает до {{max}} токенов" } }, "browser": { @@ -654,7 +663,6 @@ } }, "thinkingBudget": { - "maxTokens": "Максимум токенов", "maxThinkingTokens": "Максимум токенов на размышления" }, "validation": { diff --git a/webview-ui/src/i18n/locales/tr/settings.json b/webview-ui/src/i18n/locales/tr/settings.json index 9b90c97fb3..0765ad6d10 100644 --- a/webview-ui/src/i18n/locales/tr/settings.json +++ b/webview-ui/src/i18n/locales/tr/settings.json @@ -386,6 +386,15 @@ "placeholder": "Varsayılan: claude", "maxTokensLabel": "Maksimum Çıktı Token sayısı", "maxTokensDescription": "Claude Code yanıtları için maksimum çıktı token sayısı. Varsayılan 8000'dir." + }, + "maxOutputTokens": { + "label": "Maksimum Çıktı Belirteci", + "description": "Model çıktısı için ayrılacak maksimum belirteç sayısı. Varsayılan 8192'dir.", + "validation": { + "tooHigh": "Değer, modelin maksimum {{max}} belirteçini aşıyor", + "tooLow": "Değer en az {{min}} belirteç olmalıdır" + }, + "modelSupports": "Bu model {{max}} belirtece kadar destekler" } }, "browser": { @@ -654,7 +663,6 @@ } }, "thinkingBudget": { - "maxTokens": "Maksimum token", "maxThinkingTokens": "Maksimum düşünme tokeni" }, "validation": { diff --git a/webview-ui/src/i18n/locales/vi/settings.json b/webview-ui/src/i18n/locales/vi/settings.json index 19b574af36..49af60800f 100644 --- a/webview-ui/src/i18n/locales/vi/settings.json +++ b/webview-ui/src/i18n/locales/vi/settings.json @@ -386,6 +386,15 @@ "placeholder": "Mặc định: claude", "maxTokensLabel": "Số token đầu ra tối đa", "maxTokensDescription": "Số lượng token đầu ra tối đa cho các phản hồi của Claude Code. Mặc định là 8000." + }, + "maxOutputTokens": { + "label": "Token đầu ra tối đa", + "description": "Số lượng token tối đa để dành cho đầu ra của mô hình. Mặc định là 8192.", + "validation": { + "tooHigh": "Giá trị vượt quá tối đa của mô hình là {{max}} token", + "tooLow": "Giá trị phải ít nhất {{min}} token" + }, + "modelSupports": "Mô hình này hỗ trợ tối đa {{max}} token" } }, "browser": { @@ -654,7 +663,6 @@ } }, "thinkingBudget": { - "maxTokens": "Tokens tối đa", "maxThinkingTokens": "Tokens suy nghĩ tối đa" }, "validation": { diff --git a/webview-ui/src/i18n/locales/zh-CN/settings.json b/webview-ui/src/i18n/locales/zh-CN/settings.json index 402de1e683..feb3bd76dc 100644 --- a/webview-ui/src/i18n/locales/zh-CN/settings.json +++ b/webview-ui/src/i18n/locales/zh-CN/settings.json @@ -386,6 +386,15 @@ "placeholder": "默认：claude", "maxTokensLabel": "最大输出 Token", "maxTokensDescription": "Claude Code 响应的最大输出 Token 数量。默认为 8000。" + }, + "maxOutputTokens": { + "label": "最大输出令牌", + "description": "为模型输出保留的最大令牌数。默认值为 8192。", + "validation": { + "tooHigh": "值超过模型的最大值 {{max}} 个令牌", + "tooLow": "值必须至少为 {{min}} 个令牌" + }, + "modelSupports": "此模型支持最多 {{max}} 个令牌" } }, "browser": { @@ -654,7 +663,6 @@ } }, "thinkingBudget": { - "maxTokens": "最大Token数", "maxThinkingTokens": "最大思考Token数" }, "validation": { diff --git a/webview-ui/src/i18n/locales/zh-TW/settings.json b/webview-ui/src/i18n/locales/zh-TW/settings.json index 068e64d21f..86fb14e82f 100644 --- a/webview-ui/src/i18n/locales/zh-TW/settings.json +++ b/webview-ui/src/i18n/locales/zh-TW/settings.json @@ -386,6 +386,15 @@ "placeholder": "預設：claude", "maxTokensLabel": "最大輸出 Token", "maxTokensDescription": "Claude Code 回應的最大輸出 Token 數量。預設為 8000。" + }, + "maxOutputTokens": { + "label": "最大輸出令牌", + "description": "為模型輸出保留的最大令牌數。預設值為 8192。", + "validation": { + "tooHigh": "值超過模型的最大值 {{max}} 個令牌", + "tooLow": "值必須至少為 {{min}} 個令牌" + }, + "modelSupports": "此模型支援最多 {{max}} 個令牌" } }, "browser": { @@ -654,7 +663,6 @@ } }, "thinkingBudget": { - "maxTokens": "最大 token 數", "maxThinkingTokens": "最大思考 token 數" }, "validation": {