diff --git a/internal/providers/configs/openrouter.json b/internal/providers/configs/openrouter.json
index ec8b3fc2..741e3e7c 100644
--- a/internal/providers/configs/openrouter.json
+++ b/internal/providers/configs/openrouter.json
@@ -190,7 +190,7 @@
       "cost_per_1m_in_cached": 3.75,
       "cost_per_1m_out_cached": 0.3,
       "context_window": 200000,
-      "default_max_tokens": 32000,
+      "default_max_tokens": 64000,
       "can_reason": true,
       "reasoning_levels": [
         "low",
@@ -731,7 +731,7 @@
       "cost_per_1m_in_cached": 0,
       "cost_per_1m_out_cached": 0,
       "context_window": 1048576,
-      "default_max_tokens": 32768,
+      "default_max_tokens": 32767,
       "can_reason": true,
       "reasoning_levels": [
         "low",
@@ -1437,10 +1437,10 @@
     {
       "id": "moonshotai/kimi-k2",
       "name": "MoonshotAI: Kimi K2 0711",
-      "cost_per_1m_in": 0.5,
-      "cost_per_1m_out": 2.4,
+      "cost_per_1m_in": 0.6,
+      "cost_per_1m_out": 2.5,
       "cost_per_1m_in_cached": 0,
-      "cost_per_1m_out_cached": 0,
+      "cost_per_1m_out_cached": 0.15,
       "context_window": 131072,
       "default_max_tokens": 13107,
       "can_reason": false,
@@ -1760,9 +1760,9 @@
       "cost_per_1m_in": 0.09999999999999999,
       "cost_per_1m_out": 0.39999999999999997,
       "cost_per_1m_in_cached": 0,
-      "cost_per_1m_out_cached": 0.024999999999999998,
+      "cost_per_1m_out_cached": 0.03,
       "context_window": 1047576,
-      "default_max_tokens": 16384,
+      "default_max_tokens": 104757,
       "can_reason": false,
       "supports_attachments": true,
       "options": {}
@@ -3056,7 +3056,7 @@
       "id": "alibaba/tongyi-deepresearch-30b-a3b",
       "name": "Tongyi DeepResearch 30B A3B",
       "cost_per_1m_in": 0.09,
-      "cost_per_1m_out": 0.44999999999999996,
+      "cost_per_1m_out": 0.39999999999999997,
       "cost_per_1m_in_cached": 0,
       "cost_per_1m_out_cached": 0,
       "context_window": 131072,
@@ -3242,9 +3242,9 @@
       "cost_per_1m_in": 0.3,
       "cost_per_1m_out": 0.8999999999999999,
       "cost_per_1m_in_cached": 0,
-      "cost_per_1m_out_cached": 0,
+      "cost_per_1m_out_cached": 0.055,
       "context_window": 131072,
-      "default_max_tokens": 13107,
+      "default_max_tokens": 16384,
       "can_reason": true,
       "reasoning_levels": [
         "low",
@@ -3284,10 +3284,10 @@
     {
       "id": "x-ai/grok-3-mini",
       "name": "xAI: Grok 3 Mini",
-      "cost_per_1m_in": 0.3,
-      "cost_per_1m_out": 0.5,
+      "cost_per_1m_in": 0.6,
+      "cost_per_1m_out": 4,
       "cost_per_1m_in_cached": 0,
-      "cost_per_1m_out_cached": 0.075,
+      "cost_per_1m_out_cached": 0.15,
       "context_window": 131072,
       "default_max_tokens": 13107,
       "can_reason": true,
@@ -3303,10 +3303,10 @@
     {
       "id": "x-ai/grok-3-mini-beta",
       "name": "xAI: Grok 3 Mini Beta",
-      "cost_per_1m_in": 0.3,
-      "cost_per_1m_out": 0.5,
+      "cost_per_1m_in": 0.6,
+      "cost_per_1m_out": 4,
       "cost_per_1m_in_cached": 0,
-      "cost_per_1m_out_cached": 0.075,
+      "cost_per_1m_out_cached": 0.15,
       "context_window": 131072,
       "default_max_tokens": 13107,
       "can_reason": true,
diff --git a/internal/providers/configs/vultr.json b/internal/providers/configs/vultr.json
new file mode 100644
index 00000000..02b33293
--- /dev/null
+++ b/internal/providers/configs/vultr.json
@@ -0,0 +1,53 @@
+{
+  "name": "Vultr",
+  "id": "vultr",
+  "type": "openai-compat",
+  "api_key": "$VULTR_API_KEY",
+  "api_endpoint": "https://api.vultrinference.com/",
+  "default_large_model_id": "kimi-k2-instruct",
+  "default_small_model_id": "qwen2.5-32b-instruct",
+  "models": [
+    {
+      "id": "kimi-k2-instruct",
+      "name": "Kimi K2 Instruct",
+      "cost_per_1m_in": 1,
+      "cost_per_1m_out": 3,
+      "context_window": 262144,
+      "default_max_tokens": 8192,
+      "can_reason": true
+    },
+    {
+      "id": "llama-3.1-70b-instruct-fp8",
+      "name": "Llama 3.1 70B Instruct FP8",
+      "cost_per_1m_in": 0.35,
+      "cost_per_1m_out": 0.35,
+      "context_window": 131072,
+      "default_max_tokens": 4096
+    },
+    {
+      "id": "mistral-7b-v0.3",
+      "name": "Mistral 7B v0.3",
+      "cost_per_1m_in": 0.05,
+      "cost_per_1m_out": 0.05,
+      "context_window": 32768,
+      "default_max_tokens": 4096
+    },
+    {
+      "id": "deepseek-r1",
+      "name": "DeepSeek R1",
+      "cost_per_1m_in": 0.55,
+      "cost_per_1m_out": 2.19,
+      "context_window": 128000,
+      "default_max_tokens": 4096,
+      "can_reason": true
+    },
+    {
+      "id": "qwen2.5-32b-instruct",
+      "name": "Qwen 2.5 32B Instruct",
+      "cost_per_1m_in": 0.14,
+      "cost_per_1m_out": 0.42,
+      "context_window": 32768,
+      "default_max_tokens": 4096
+    }
+  ]
+}
diff --git a/internal/providers/providers.go b/internal/providers/providers.go
index d5299c49..5cf74309 100644
--- a/internal/providers/providers.go
+++ b/internal/providers/providers.go
@@ -63,6 +63,8 @@ var aiHubMixConfig []byte
 //go:embed configs/kimi.json
 var kimiCodingConfig []byte
 
+//go:embed configs/vultr.json
+var vultrConfig []byte
 //go:embed configs/copilot.json
 var copilotConfig []byte
 
@@ -87,6 +89,7 @@ var providerRegistry = []ProviderFunc{
 	deepSeekProvider,
 	huggingFaceProvider,
 	aiHubMixProvider,
+	vultrProvider,
 	syntheticProvider,
 	copilotProvider,
 }
@@ -181,6 +184,8 @@ func kimiCodingProvider() catwalk.Provider {
 	return loadProviderFromConfig(kimiCodingConfig)
 }
 
+func vultrProvider() catwalk.Provider {
+	return loadProviderFromConfig(vultrConfig)
 func copilotProvider() catwalk.Provider {
 	return loadProviderFromConfig(copilotConfig)
 }