diff --git a/src/llama-chat.cpp b/src/llama-chat.cpp
index 46d43c58ef4d4..d12743e6b9a0c 100644
--- a/src/llama-chat.cpp
+++ b/src/llama-chat.cpp
@@ -35,6 +35,7 @@ static const std::map<std::string, llm_chat_template> LLM_CHAT_TEMPLATES = {
     { "mistral-v3",        LLM_CHAT_TEMPLATE_MISTRAL_V3        },
     { "mistral-v3-tekken", LLM_CHAT_TEMPLATE_MISTRAL_V3_TEKKEN },
     { "mistral-v7",        LLM_CHAT_TEMPLATE_MISTRAL_V7        },
+    { "mistral-v7-tekken", LLM_CHAT_TEMPLATE_MISTRAL_V7_TEKKEN },
     { "phi3",              LLM_CHAT_TEMPLATE_PHI_3             },
     { "phi4",              LLM_CHAT_TEMPLATE_PHI_4             },
     { "falcon3",           LLM_CHAT_TEMPLATE_FALCON_3          },
@@ -202,19 +203,20 @@ int32_t llm_chat_apply_template(
         if (add_ass) {
             ss << "<|im_start|>assistant\n";
         }
-    } else if (tmpl == LLM_CHAT_TEMPLATE_MISTRAL_V7) {
+    } else if (tmpl == LLM_CHAT_TEMPLATE_MISTRAL_V7 || tmpl == LLM_CHAT_TEMPLATE_MISTRAL_V7_TEKKEN) {
         // Official mistral 'v7' template
         // See: https://huggingface.co/mistralai/Mistral-Large-Instruct-2411#basic-instruct-template-v7
+        //      https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503#basic-instruct-template-v7-tekken
+        const char * trailing_space = tmpl == LLM_CHAT_TEMPLATE_MISTRAL_V7 ? " " : "";
         for (auto message : chat) {
             std::string role(message->role);
             std::string content(message->content);
             if (role == "system") {
-                ss << "[SYSTEM_PROMPT] " << content << "[/SYSTEM_PROMPT]";
+                ss << "[SYSTEM_PROMPT]" << trailing_space << content << "[/SYSTEM_PROMPT]";
             } else if (role == "user") {
-                ss << "[INST] " << content << "[/INST]";
-            }
-            else {
-                ss << " " << content << "</s>";
+                ss << "[INST]" << trailing_space << content << "[/INST]";
+            } else {
+                ss << trailing_space << content << "</s>";
             }
         }
     } else if (tmpl == LLM_CHAT_TEMPLATE_MISTRAL_V1
diff --git a/src/llama-chat.h b/src/llama-chat.h
index 3f5843466d044..db24ade21e2ad 100644
--- a/src/llama-chat.h
+++ b/src/llama-chat.h
@@ -14,6 +14,7 @@ enum llm_chat_template {
     LLM_CHAT_TEMPLATE_MISTRAL_V3,
     LLM_CHAT_TEMPLATE_MISTRAL_V3_TEKKEN,
     LLM_CHAT_TEMPLATE_MISTRAL_V7,
+    LLM_CHAT_TEMPLATE_MISTRAL_V7_TEKKEN,
     LLM_CHAT_TEMPLATE_PHI_3,
     LLM_CHAT_TEMPLATE_PHI_4,
     LLM_CHAT_TEMPLATE_FALCON_3,
diff --git a/src/llama-model.cpp b/src/llama-model.cpp
index 3ca265be8dca4..e8b78c1d00296 100644
--- a/src/llama-model.cpp
+++ b/src/llama-model.cpp
@@ -13387,6 +13387,14 @@ const char * llama_model_chat_template(const llama_model * model, const char * n
         : LLM_KV(model->arch)(LLM_KV_TOKENIZER_CHAT_TEMPLATE);
     const auto & it = model->gguf_kv.find(key);
     if (it == model->gguf_kv.end()) {
+        // one-off fix for very popular models (so we are not flooded with issues)
+        // do not extend this list unless absolutely necessary
+        // Mistral-Small-2503 does not have built-in chat template
+        llama_vocab_pre_type pre_type = model->vocab.get_pre_type();
+        if (pre_type == LLAMA_VOCAB_PRE_TYPE_TEKKEN && model->layers.size() == 40) {
+            return "mistral-v7-tekken";
+        }
+
         return nullptr;
     }
 
diff --git a/tools/mtmd/README.md b/tools/mtmd/README.md
index b97b9e8c54367..20e7696cefd8e 100644
--- a/tools/mtmd/README.md
+++ b/tools/mtmd/README.md
@@ -46,7 +46,7 @@ llama-mtmd-cli -hf ggml-org/Qwen2.5-VL-32B-Instruct-GGUF
 llama-mtmd-cli -hf ggml-org/Qwen2.5-VL-72B-Instruct-GGUF
 
 # Mistral Small 3.1 24B (IQ2_M quantization)
-llama-mtmd-cli -hf ggml-org/Mistral-Small-3.1-24B-Instruct-2503-GGUF --chat-template mistral-v7
+llama-mtmd-cli -hf ggml-org/Mistral-Small-3.1-24B-Instruct-2503-GGUF
 ```
 
 ## How it works and what is `mmproj`?