diff --git a/src/config/aiModels/stepfun.ts b/src/config/aiModels/stepfun.ts
index 93c94d33b4657..bba33e2bbf1ec 100644
--- a/src/config/aiModels/stepfun.ts
+++ b/src/config/aiModels/stepfun.ts
@@ -96,6 +96,41 @@ const stepfunChatModels: AIChatModelCard[] = [
     },
     type: 'chat',
   },
+  {
+    abilities: {
+      functionCall: true,
+    },
+    contextWindowTokens: 8000,
+    description: 
+      '基于新一代自研Attention架构MFA的极速大模型，用极低成本达到和step1类似的效果，同时保持了更高的吞吐和更快响应时延。能够处理通用任务，在代码能力上具备特长。',
+    displayName: 'Step 2 Mini',
+    enabled: true,
+    id: 'step-2-mini',
+      pricing: {
+      currency: 'CNY',
+      input: 1,
+      output: 2,
+    },
+    releasedAt: '2025-01-14',
+    type: 'chat',
+  },
+  {
+    abilities: {
+      functionCall: true,
+    },
+    contextWindowTokens: 16_000,
+    description: 'step-2模型的实验版本，包含最新的特性，滚动更新中。不推荐在正式生产环境使用。',
+    displayName: 'Step 2 16K Exp',
+    enabled: true,
+    id: 'step-2-16k',
+    pricing: {
+      currency: 'CNY',
+      input: 38,
+      output: 120,
+    },
+    releasedAt: '2025-01-15',
+    type: 'chat',
+  },
   {
     abilities: {
       functionCall: true,
@@ -121,7 +156,6 @@ const stepfunChatModels: AIChatModelCard[] = [
     contextWindowTokens: 32_000,
     description: '支持视觉输入，增强多模态交互体验。',
     displayName: 'Step 1V 32K',
-    enabled: true,
     id: 'step-1v-32k',
     pricing: {
       currency: 'CNY',
@@ -130,6 +164,24 @@ const stepfunChatModels: AIChatModelCard[] = [
     },
     type: 'chat',
   },
+  {
+    abilities: {
+      functionCall: true,
+      vision: true,
+    },
+    contextWindowTokens: 32_000,
+    description: '该模型拥有强大的图像理解能力。相比于 step-1v 系列模型，拥有更强的视觉性能。',
+    displayName: 'Step 1o Vision 32K',
+    enabled: true,
+    id: 'step-1o-vision-32k',
+    pricing: {
+      currency: 'CNY',
+      input: 15,
+      output: 70,
+    },
+    releasedAt: '2025-01-22',
+    type: 'chat',
+  },
   {
     abilities: {
       vision: true,
diff --git a/src/config/modelProviders/stepfun.ts b/src/config/modelProviders/stepfun.ts
index 9961be102bba1..d78928fbee03b 100644
--- a/src/config/modelProviders/stepfun.ts
+++ b/src/config/modelProviders/stepfun.ts
@@ -81,6 +81,32 @@ const Stepfun: ModelProviderCard = {
         output: 120,
       },
     },
+    {
+      contextWindowTokens: 8000,
+      description: 
+        '基于新一代自研Attention架构MFA的极速大模型，用极低成本达到和step1类似的效果，同时保持了更高的吞吐和更快响应时延。能够处理通用任务，在代码能力上具备特长。',
+      displayName: 'Step 2 Mini',
+      enabled: true,
+      functionCall: true,
+      id: 'step-2-mini',
+      pricing: {
+        currency: 'CNY',
+        input: 1,
+        output: 2,
+      },
+    },
+    {
+      contextWindowTokens: 16_000,
+      description: 'step-2模型的实验版本，包含最新的特性，滚动更新中。不推荐在正式生产环境使用。',
+      displayName: 'Step 2 16K Exp',
+      functionCall: true,
+      id: 'step-2-16k-exp',
+      pricing: {
+        currency: 'CNY',
+        input: 38,
+        output: 120,
+      },
+    },
     {
       contextWindowTokens: 8000,
       description: '小型视觉模型，适合基本的图文任务。',
@@ -99,7 +125,6 @@ const Stepfun: ModelProviderCard = {
       contextWindowTokens: 32_000,
       description: '支持视觉输入，增强多模态交互体验。',
       displayName: 'Step 1V 32K',
-      enabled: true,
       functionCall: true,
       id: 'step-1v-32k',
       pricing: {
@@ -109,6 +134,19 @@ const Stepfun: ModelProviderCard = {
       },
       vision: true,
     },
+    {
+      contextWindowTokens: 32_000,
+      description: '该模型拥有强大的图像理解能力。相比于 step-1v 系列模型，拥有更强的视觉性能。',
+      displayName: 'Step 1o Vision 32K',
+      enabled: true,
+      id: 'step-1o-vision-32k',
+      pricing: {
+        currency: 'CNY',
+        input: 15,
+        output: 70,
+      },
+      vision: true,
+    },
     {
       contextWindowTokens: 32_000,
       description: '该模型拥有强大的视频理解能力。',
@@ -123,7 +161,7 @@ const Stepfun: ModelProviderCard = {
       vision: true,
     },
   ],
-  checkModel: 'step-1-flash',
+  checkModel: 'step-2-mini',
   description:
     '阶级星辰大模型具备行业领先的多模态及复杂推理能力，支持超长文本理解和强大的自主调度搜索引擎功能。',
   // after test, currently https://api.stepfun.com/v1/chat/completions has the CORS issue