c221094a7fa11dfea8ad5ba4bc90f7fcdb748086

 1diff --git a/dot_config/llama-server/models.ini b/dot_config/llama-server/models.ini
 2index 4bdacf387ec534215ff40ed1e323414c1b7db898..f3d74905795e4da0ec2baa2cce2a4812f278515b 100644
 3--- a/dot_config/llama-server/models.ini
 4+++ b/dot_config/llama-server/models.ini
 5@@ -15,9 +15,9 @@ jinja = true
 6 temp = 0.15
 7 parallel = 1
 8 
 9-[unsloth/Qwen3.6-27B-MTP-GGUF:Q4_K_M]
10-hf = unsloth/Qwen3.6-27B-MTP-GGUF:Q4_K_M
11-alias = qwen3.6-27b
12+[unsloth/Qwen3.8-27B-GGUF:Q4_K_XL]
13+hf = unsloth/Qwen3.8-27B-GGUF:UD-Q4_K_XL
14+alias = qwen3.8-27b
15 n-gpu-layers = 99
16 flash-attn = on
17 jinja = true
18@@ -31,6 +31,7 @@ spec-type = draft-mtp
19 spec-draft-n-max = 2
20 repeat-penalty = 1.0
21 reasoning = on
22+chat-template-kwargs = {"reasoning_effort":"medium"}
23 
24 [unsloth/Qwen3.6-35B-A3B-GGUF:Q4_K_S]
25 hf = unsloth/Qwen3.6-35B-A3B-GGUF:UD-Q4_K_S