diff --git a/src/onboard/wizard.rs b/src/onboard/wizard.rs
index 00d87b980..7ec287d3d 100644
--- a/src/onboard/wizard.rs
+++ b/src/onboard/wizard.rs
@@ -783,10 +783,15 @@ fn allows_unauthenticated_model_fetch(provider_name: &str) -> bool {
 }
 
 /// Pick a sensible default model for the given provider.
-const MINIMAX_ONBOARD_MODELS: [(&str, &str); 5] = [
-    ("MiniMax-M2.5", "MiniMax M2.5 (latest, recommended)"),
+const MINIMAX_ONBOARD_MODELS: [(&str, &str); 7] = [
+    (
+        "MiniMax-M2.7",
+        "MiniMax M2.7 (latest flagship, recommended)",
+    ),
+    ("MiniMax-M2.7-highspeed", "MiniMax M2.7 High-Speed (faster)"),
+    ("MiniMax-M2.5", "MiniMax M2.5 (stable)"),
     ("MiniMax-M2.5-highspeed", "MiniMax M2.5 High-Speed (faster)"),
-    ("MiniMax-M2.1", "MiniMax M2.1 (stable)"),
+    ("MiniMax-M2.1", "MiniMax M2.1 (previous gen)"),
     ("MiniMax-M2.1-highspeed", "MiniMax M2.1 High-Speed (faster)"),
     ("MiniMax-M2", "MiniMax M2 (legacy)"),
 ];
@@ -803,12 +808,12 @@ fn default_model_for_provider(provider: &str) -> String {
         "xai" => "grok-4-1-fast-reasoning".into(),
         "perplexity" => "sonar-pro".into(),
         "fireworks" => "accounts/fireworks/models/llama-v3p3-70b-instruct".into(),
-        "novita" => "minimax/minimax-m2.5".into(),
+        "novita" => "minimax/minimax-m2.7".into(),
         "together-ai" => "meta-llama/Llama-3.3-70B-Instruct-Turbo".into(),
         "cohere" => "command-a-03-2025".into(),
         "moonshot" => "kimi-k2.5".into(),
         "glm" | "zai" => "glm-5".into(),
-        "minimax" => "MiniMax-M2.5".into(),
+        "minimax" => "MiniMax-M2.7".into(),
         "qwen" => "qwen-plus".into(),
         "qwen-code" => "qwen3-coder-plus".into(),
         "ollama" => "llama3.2".into(),
@@ -997,10 +1002,16 @@ fn curated_models_for_provider(provider_name: &str) -> Vec<(String, String)> {
                 "Mixtral 8x22B".to_string(),
             ),
         ],
-        "novita" => vec![(
-            "minimax/minimax-m2.5".to_string(),
-            "MiniMax M2.5".to_string(),
-        )],
+        "novita" => vec![
+            (
+                "minimax/minimax-m2.7".to_string(),
+                "MiniMax M2.7 (latest flagship)".to_string(),
+            ),
+            (
+                "minimax/minimax-m2.5".to_string(),
+                "MiniMax M2.5".to_string(),
+            ),
+        ],
         "together-ai" => vec![
             (
                 "meta-llama/Llama-3.3-70B-Instruct-Turbo".to_string(),
@@ -1065,9 +1076,17 @@ fn curated_models_for_provider(provider_name: &str) -> Vec<(String, String)> {
             ),
         ],
         "minimax" => vec![
+            (
+                "MiniMax-M2.7".to_string(),
+                "MiniMax M2.7 (latest flagship)".to_string(),
+            ),
+            (
+                "MiniMax-M2.7-highspeed".to_string(),
+                "MiniMax M2.7 High-Speed (fast)".to_string(),
+            ),
             (
                 "MiniMax-M2.5".to_string(),
-                "MiniMax M2.5 (latest flagship)".to_string(),
+                "MiniMax M2.5 (stable)".to_string(),
             ),
             (
                 "MiniMax-M2.5-highspeed".to_string(),
@@ -1075,7 +1094,7 @@ fn curated_models_for_provider(provider_name: &str) -> Vec<(String, String)> {
             ),
             (
                 "MiniMax-M2.1".to_string(),
-                "MiniMax M2.1 (strong coding/reasoning)".to_string(),
+                "MiniMax M2.1 (previous gen)".to_string(),
             ),
         ],
         "qwen" => vec![
@@ -1621,7 +1640,7 @@ fn fetch_live_models_for_provider(
                     "qwen3-coder-next:cloud".to_string(),
                     "qwen3-coder:480b:cloud".to_string(),
                     "kimi-k2.5:cloud".to_string(),
-                    "minimax-m2.5:cloud".to_string(),
+                    "minimax-m2.7:cloud".to_string(),
                     "deepseek-v3.1:671b:cloud".to_string(),
                 ]
             } else {
@@ -6651,7 +6670,7 @@ mod tests {
         assert_eq!(default_model_for_provider("qwen-intl"), "qwen-plus");
         assert_eq!(default_model_for_provider("qwen-code"), "qwen3-coder-plus");
         assert_eq!(default_model_for_provider("glm-cn"), "glm-5");
-        assert_eq!(default_model_for_provider("minimax-cn"), "MiniMax-M2.5");
+        assert_eq!(default_model_for_provider("minimax-cn"), "MiniMax-M2.7");
         assert_eq!(default_model_for_provider("zai-cn"), "glm-5");
         assert_eq!(default_model_for_provider("gemini"), "gemini-2.5-pro");
         assert_eq!(default_model_for_provider("google"), "gemini-2.5-pro");