Skip to content

fix: apply mergeToolResultText only for Mistral family models in OpenAI provider #10716

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
roomote wants to merge 1 commit into from
+22 −4
Closed

fix: apply mergeToolResultText only for Mistral family models in OpenAI provider #10716

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
26 changes: 22 additions & 4 deletions src/api/providers/openai.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -92,6 +92,11 @@ export class OpenAiHandler extends BaseProvider implements SingleCompletionHandl
const isAzureAiInference = this._isAzureAiInference(modelUrl)
const deepseekReasoner = modelId.includes("deepseek-reasoner") || enabledR1Format

// Check if model is Mistral/Devstral family to apply strict message ordering fix
// Mistral models reject "user" role immediately after "tool" role
const modelIdLower = modelId.toLowerCase()
const isMistralFamily = modelIdLower.includes("mistral") || modelIdLower.includes("devstral")

if (modelId.includes("o1") || modelId.includes("o3") || modelId.includes("o4")) {
yield* this.handleO3FamilyMessage(modelId, systemPrompt, messages, metadata)
return
Expand Down Expand Up @@ -122,7 +127,10 @@ export class OpenAiHandler extends BaseProvider implements SingleCompletionHandl
}
}

convertedMessages = [systemMessage, ...convertToOpenAiMessages(messages)]
convertedMessages = [
systemMessage,
...convertToOpenAiMessages(messages, isMistralFamily ? { mergeToolResultText: true } : undefined),
]

if (modelInfo.supportsPromptCache) {
// Note: the following logic is copied from openrouter:
Expand Down Expand Up @@ -229,7 +237,13 @@ export class OpenAiHandler extends BaseProvider implements SingleCompletionHandl
model: modelId,
messages: deepseekReasoner
? convertToR1Format([{ role: "user", content: systemPrompt }, ...messages])
: [systemMessage, ...convertToOpenAiMessages(messages)],
: [
systemMessage,
...convertToOpenAiMessages(
messages,
isMistralFamily ? { mergeToolResultText: true } : undefined,
),
],
...(metadata?.tools && { tools: this.convertToolsForOpenAI(metadata.tools) }),
...(metadata?.tool_choice && { tool_choice: metadata.tool_choice }),
...(metadata?.toolProtocol === "native" &&
Expand Down Expand Up @@ -341,6 +355,10 @@ export class OpenAiHandler extends BaseProvider implements SingleCompletionHandl
const modelInfo = this.getModel().info
const methodIsAzureAiInference = this._isAzureAiInference(this.options.openAiBaseUrl)

// Check if model is Mistral/Devstral family for strict message ordering fix
const modelIdLower = modelId.toLowerCase()
const isMistralFamily = modelIdLower.includes("mistral") || modelIdLower.includes("devstral")

if (this.options.openAiStreamingEnabled ?? true) {
const isGrokXAI = this._isGrokXAI(this.options.openAiBaseUrl)

Expand All @@ -351,7 +369,7 @@ export class OpenAiHandler extends BaseProvider implements SingleCompletionHandl
role: "developer",
content: `Formatting re-enabled\n${systemPrompt}`,
},
...convertToOpenAiMessages(messages),
...convertToOpenAiMessages(messages, isMistralFamily ? { mergeToolResultText: true } : undefined),
],
stream: true,
...(isGrokXAI ? {} : { stream_options: { include_usage: true } }),
Expand Down Expand Up @@ -389,7 +407,7 @@ export class OpenAiHandler extends BaseProvider implements SingleCompletionHandl
role: "developer",
content: `Formatting re-enabled\n${systemPrompt}`,
},
...convertToOpenAiMessages(messages),
...convertToOpenAiMessages(messages, isMistralFamily ? { mergeToolResultText: true } : undefined),
],
reasoning_effort: modelInfo.reasoningEffort as "low" | "medium" | "high" | undefined,
temperature: undefined,
Expand Down
Loading