💄 style: add pricing info for Azure GPT-5 series models (#9833)

* 💰 feat: add pricing info for Azure GPT-5 series models Added comprehensive pricing information for all Azure OpenAI GPT-5 series models: - GPT-5 Pro: $15/$120 per million tokens (input/output) - GPT-5 Codex: $1.25/$10 per million tokens with cache support - GPT-5: $1.25/$10 per million tokens with cache support - GPT-5 Mini: $0.25/$2 per million tokens with cache support - GPT-5 Nano: $0.05/$0.4 per million tokens with cache support - GPT-5 Chat: $1.25/$10 per million tokens with cache support Pricing aligns with OpenAI official rates and includes cache read pricing where applicable. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> * 🔥 feat: remove Azure GPT-5 series models Removed all GPT-5 series model configurations from Azure provider: - GPT-5 Pro - GPT-5 Codex - GPT-5 - GPT-5 Mini - GPT-5 Nano - GPT-5 Chat These models are not yet officially released by Azure OpenAI. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> --------- Co-authored-by: Zhao <zhaokm@ZhaodeMacBook-Pro.local> Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: Zhao <zhaokm@842f57833757mbpnetwork.lan>
2026-03-27 13:29:15 +07:00 · 2025-10-31 01:02:48 +08:00
parent 1b8a981e5b
commit 39a80c5604
3 changed files with 160 additions and 1 deletions
--- a/packages/const/src/models.ts
+++ b/packages/const/src/models.ts
@@ -35,6 +35,9 @@ export const responsesAPIModels = new Set([
  'codex-mini-latest',
  'computer-use-preview',
  'computer-use-preview-2025-03-11',
+  'gpt-5-codex',
+  'gpt-5-pro',
+  'gpt-5-pro-2025-10-06',
 ]);

 /**
--- a/packages/model-bank/src/aiModels/azure.ts
+++ b/packages/model-bank/src/aiModels/azure.ts
@@ -1,6 +1,161 @@
 import { AIChatModelCard, AIImageModelCard } from '../types/aiModel';

 const azureChatModels: AIChatModelCard[] = [
+  {
+    abilities: {
+      functionCall: true,
+      reasoning: true,
+      structuredOutput: true,
+      vision: true,
+    },
+    config: {
+      deploymentName: 'gpt-5-pro',
+    },
+    contextWindowTokens: 400_000,
+    description:
+      'GPT-5 Pro 是 GPT-5 系列的高级版本，具备增强的推理能力。支持结构化输出、函数调用和文本/图像处理，适用于复杂的专业任务。',
+    displayName: 'GPT-5 Pro',
+    enabled: true,
+    id: 'gpt-5-pro',
+    maxOutput: 128_000,
+    pricing: {
+      units: [
+        { name: 'textInput', rate: 15, strategy: 'fixed', unit: 'millionTokens' },
+        { name: 'textOutput', rate: 120, strategy: 'fixed', unit: 'millionTokens' },
+      ],
+    },
+    releasedAt: '2025-10-06',
+    type: 'chat',
+  },
+  {
+    abilities: {
+      functionCall: true,
+      structuredOutput: true,
+    },
+    config: {
+      deploymentName: 'gpt-5-codex',
+    },
+    contextWindowTokens: 400_000,
+    description:
+      'GPT-5 Codex 专为编程任务优化，针对 Codex CLI 和 VS Code 扩展进行了优化。支持结构化输出和函数调用，适用于代码生成和分析。',
+    displayName: 'GPT-5 Codex',
+    enabled: true,
+    id: 'gpt-5-codex',
+    maxOutput: 128_000,
+    pricing: {
+      units: [
+        { name: 'textInput', rate: 1.25, strategy: 'fixed', unit: 'millionTokens' },
+        { name: 'textOutput', rate: 10, strategy: 'fixed', unit: 'millionTokens' },
+        { name: 'textInput_cacheRead', rate: 0.125, strategy: 'fixed', unit: 'millionTokens' },
+      ],
+    },
+    releasedAt: '2025-09-11',
+    type: 'chat',
+  },
+  {
+    abilities: {
+      functionCall: true,
+      reasoning: true,
+      structuredOutput: true,
+      vision: true,
+    },
+    config: {
+      deploymentName: 'gpt-5',
+    },
+    contextWindowTokens: 400_000,
+    description:
+      'GPT-5 是 OpenAI 最新的旗舰模型，具备卓越的推理能力。支持文本和图像输入，结构化输出和并行工具调用，适用于需要深度理解和分析的复杂任务。',
+    displayName: 'GPT-5',
+    enabled: true,
+    id: 'gpt-5',
+    maxOutput: 128_000,
+    pricing: {
+      units: [
+        { name: 'textInput', rate: 1.25, strategy: 'fixed', unit: 'millionTokens' },
+        { name: 'textOutput', rate: 10, strategy: 'fixed', unit: 'millionTokens' },
+        { name: 'textInput_cacheRead', rate: 0.125, strategy: 'fixed', unit: 'millionTokens' },
+      ],
+    },
+    releasedAt: '2025-08-07',
+    type: 'chat',
+  },
+  {
+    abilities: {
+      functionCall: true,
+      reasoning: true,
+      structuredOutput: true,
+      vision: true,
+    },
+    config: {
+      deploymentName: 'gpt-5-mini',
+    },
+    contextWindowTokens: 400_000,
+    description:
+      'GPT-5 Mini 提供与 GPT-5 相似的能力，但更加高效和经济。支持推理、函数调用和视觉功能，适合大规模部署和对成本敏感的应用场景。',
+    displayName: 'GPT-5 Mini',
+    enabled: true,
+    id: 'gpt-5-mini',
+    maxOutput: 128_000,
+    pricing: {
+      units: [
+        { name: 'textInput', rate: 0.25, strategy: 'fixed', unit: 'millionTokens' },
+        { name: 'textOutput', rate: 2, strategy: 'fixed', unit: 'millionTokens' },
+        { name: 'textInput_cacheRead', rate: 0.025, strategy: 'fixed', unit: 'millionTokens' },
+      ],
+    },
+    releasedAt: '2025-08-07',
+    type: 'chat',
+  },
+  {
+    abilities: {
+      functionCall: true,
+      reasoning: true,
+      structuredOutput: true,
+      vision: true,
+    },
+    config: {
+      deploymentName: 'gpt-5-nano',
+    },
+    contextWindowTokens: 400_000,
+    description:
+      'GPT-5 Nano 是 GPT-5 系列中最小、最快的版本。在保持核心能力的同时，提供超低延迟和成本效益，适合边缘计算和实时应用。',
+    displayName: 'GPT-5 Nano',
+    enabled: true,
+    id: 'gpt-5-nano',
+    maxOutput: 128_000,
+    pricing: {
+      units: [
+        { name: 'textInput', rate: 0.05, strategy: 'fixed', unit: 'millionTokens' },
+        { name: 'textOutput', rate: 0.4, strategy: 'fixed', unit: 'millionTokens' },
+        { name: 'textInput_cacheRead', rate: 0.005, strategy: 'fixed', unit: 'millionTokens' },
+      ],
+    },
+    releasedAt: '2025-08-07',
+    type: 'chat',
+  },
+  {
+    abilities: {
+      vision: true,
+    },
+    config: {
+      deploymentName: 'gpt-5-chat',
+    },
+    contextWindowTokens: 128_000,
+    description:
+      'GPT-5 Chat 专为对话场景优化的预览版本。支持文本和图像输入，仅输出文本，适用于聊天机器人和对话式AI应用。',
+    displayName: 'GPT-5 Chat',
+    id: 'gpt-5-chat',
+    maxOutput: 16_384,
+    pricing: {
+      units: [
+        { name: 'textInput', rate: 1.25, strategy: 'fixed', unit: 'millionTokens' },
+        { name: 'textOutput', rate: 10, strategy: 'fixed', unit: 'millionTokens' },
+        { name: 'textInput_cacheRead', rate: 0.125, strategy: 'fixed', unit: 'millionTokens' },
+      ],
+    },
+    releasedAt: '2025-08-07',
+    type: 'chat',
+  },
  {
    abilities: {
      functionCall: true,
--- a/packages/model-runtime/src/providers/azureOpenai/index.ts
+++ b/packages/model-runtime/src/providers/azureOpenai/index.ts
@@ -51,7 +51,8 @@ export class LobeAzureOpenAI implements LobeRuntimeAI {
      ...message,
      role:
        // Convert 'system' role to 'user' or 'developer' based on the model
-        (model.includes('o1') || model.includes('o3')) && message.role === 'system'
+        (model.includes('o1') || model.includes('o3') || model.includes('gpt-5')) &&
+        message.role === 'system'
          ? [...systemToUserModels].some((sub) => model.includes(sub))
            ? 'user'
            : 'developer'