Northern_Lights/cherry-studio
1
0
Fork 0
mirror of https://github.com/CherryHQ/cherry-studio.git synced 2025-12-29 23:12:38 +08:00
Code Issues Actions 3 Packages Projects Releases Wiki Activity

fix: respect enableMaxTokens setting when maxTokens is not configured (#11438)

Browse Source 
* fix: respect enableMaxTokens setting when maxTokens is not configured

When enableMaxTokens is disabled, getMaxTokens() should return undefined
to let the API use its own default value, instead of forcing 4096 tokens.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

* fix(modelParameters): handle max tokens when feature is disabled

Check if max tokens feature is enabled before returning undefined to ensure proper API behavior

---------

Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: icarus <eurfelux@gmail.com>
This commit is contained in:
fullex 2025-11-25 11:12:50 +08:00 committed by GitHub
parent 1992363580
commit 0004a8cafe
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 10 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
src/renderer/src/aiCore/prepareParams/modelParameters.ts
View File

@ -3,7 +3,6 @@
* 处理温度、TopP、超时等基础参数的获取逻辑
*/
import { DEFAULT_MAX_TOKENS } from '@renderer/config/constant'
import {
isClaude45ReasoningModel,
isClaudeReasoningModel,
@ -73,11 +72,19 @@ export function getTimeout(model: Model): number {
export function getMaxTokens(assistant: Assistant, model: Model): number | undefined {
// NOTE: ai-sdk会把maxToken和budgetToken加起来
let { maxTokens = DEFAULT_MAX_TOKENS } = getAssistantSettings(assistant)
const assistantSettings = getAssistantSettings(assistant)
const enabledMaxTokens = assistantSettings.enableMaxTokens ?? false
let maxTokens = assistantSettings.maxTokens
// If user hasn't enabled enableMaxTokens, return undefined to let the API use its default value.
// Note: Anthropic API requires max_tokens, but that's handled by the Anthropic client with a fallback.
if (!enabledMaxTokens || maxTokens === undefined) {
return undefined
}
const provider = getProviderByModel(model)
if (isSupportedThinkingTokenClaudeModel(model) && ['anthropic', 'aws-bedrock'].includes(provider.type)) {
const { reasoning_effort: reasoningEffort } = getAssistantSettings(assistant)
const { reasoning_effort: reasoningEffort } = assistantSettings
const budget = getAnthropicThinkingBudget(maxTokens, reasoningEffort, model.id)
if (budget) {
maxTokens -= budget