prefill cannot be used with thinking models

matteoserva · matteoserva · commit 226e37d8e023 · 2025-06-08T14:25:10.000+02:00
diff --git a/tools/server/utils.hpp b/tools/server/utils.hpp
@@ -772,7 +772,7 @@ static json oaicompat_chat_params_parse(
         /* TODO: test this properly */
         inputs.reasoning_format = COMMON_REASONING_FORMAT_NONE;
 
-        if (inputs.chat_template_kwargs.find("enable_thinking") != inputs.chat_template_kwargs.end()) {
+        if (inputs.enable_thinking || inputs.chat_template_kwargs.find("enable_thinking") != inputs.chat_template_kwargs.end()) {
             throw std::runtime_error("Assistant response prefill is incompatible with enable_thinking.");
         }
 

Original file line number	Diff line number	Diff line change
`@@ -772,7 +772,7 @@ static json oaicompat_chat_params_parse(`
`772`	`772`	`/* TODO: test this properly */`
`773`	`773`	`inputs.reasoning_format = COMMON_REASONING_FORMAT_NONE;`
`774`	`774`
`775`		`- if (inputs.chat_template_kwargs.find("enable_thinking") != inputs.chat_template_kwargs.end()) {`
	`775`	`+ if (inputs.enable_thinking \|\| inputs.chat_template_kwargs.find("enable_thinking") != inputs.chat_template_kwargs.end()) {`
`776`	`776`	`throw std::runtime_error("Assistant response prefill is incompatible with enable_thinking.");`
`777`	`777`	`}`
`778`	`778`