From d7ff074c87ecacd57d5760e2f678866ba9fe7149 Mon Sep 17 00:00:00 2001
From: Berk Idem <55372926+berkidem@users.noreply.github.com>
Date: Fri, 10 Apr 2026 05:49:14 -0400
Subject: [PATCH] common : enable reasoning budget sampler for gemma4 (#21697)

* fix: enable reasoning budget sampler for gemma4

Add thinking_start_tag and thinking_end_tag to
common_chat_params_init_gemma4(). Without these, the reasoning
budget sampler never activates for gemma4.

Make the newline after "thought" optional in the PEG parser to
handle budget=0 (sampler forces end tag before the newline).

Add test case for empty thinking block.

Fixes #21487

* use p.space() instead of p.optional(p.literal("\n")) in gemma4 thought parser
---
 common/chat.cpp     | 8 +++++---
 tests/test-chat.cpp | 7 +++++++
 2 files changed, 12 insertions(+), 3 deletions(-)

diff --git a/common/chat.cpp b/common/chat.cpp
index f79fe703db..03cd9e25f4 100644
--- a/common/chat.cpp
+++ b/common/chat.cpp
@@ -1083,7 +1083,9 @@ static common_chat_params common_chat_params_init_gemma4(const common_chat_templ
 
     data.prompt            = common_chat_template_direct_apply_impl(tmpl, inputs);
     data.format            = COMMON_CHAT_FORMAT_PEG_GEMMA4;
-    data.supports_thinking = true;
+    data.supports_thinking  = true;
+    data.thinking_start_tag = "<|channel>thought";
+    data.thinking_end_tag   = "<channel|>";
 
     data.preserved_tokens = {
         "<|channel>",
@@ -1102,9 +1104,9 @@ static common_chat_params common_chat_params_init_gemma4(const common_chat_templ
         auto start = p.rule("start", p.prefix(inputs.generation_prompt, "<|channel>"));
 
         if (extract_reasoning) {
-            p.rule("thought", p.literal("<|channel>thought\n") + p.reasoning(p.until("<channel|>")) + p.literal("<channel|>"));
+            p.rule("thought", p.literal("<|channel>thought") + p.space() + p.reasoning(p.until("<channel|>")) + p.literal("<channel|>"));
         } else {
-            p.rule("thought", p.content(p.literal("<|channel>thought\n") + p.until("<channel|>") + p.literal("<channel|>")));
+            p.rule("thought", p.content(p.literal("<|channel>thought") + p.space() + p.until("<channel|>") + p.literal("<channel|>")));
         }
 
         auto thought = (p.peek(p.literal("<|channel>")) + p.ref("thought")) | p.negate(p.literal("<|channel>"));
diff --git a/tests/test-chat.cpp b/tests/test-chat.cpp
index f8b2b584e2..1d3a77bc05 100644
--- a/tests/test-chat.cpp
+++ b/tests/test-chat.cpp
@@ -1988,6 +1988,13 @@ static void test_template_output_peg_parsers(bool detailed_debug) {
             .expect(message_assist_thoughts)
             .run();
 
+        // Empty reasoning (budget=0: sampler forces end tag before newline)
+        tst.test(
+                "<|channel>thought<channel|>Hello, world!\nWhat's up?")
+            .reasoning_format(COMMON_REASONING_FORMAT_AUTO)
+            .expect(simple_assist_msg("Hello, world!\nWhat's up?", ""))
+            .run();
+
         // Reasoning and content with reasoning_format = none
         tst.test(
                 "<|channel>thought\nI'm\nthinking<channel|>Hello, world!\nWhat's up?")