icebear0828 Claude Opus 4.6 commited on
Commit
43aa0a0
·
1 Parent(s): f678c8b

docs: update model catalog and changelog for 2026-03-10 changes

Browse files

- CHANGELOG: add cached_tokens passthrough (#55/#58), model-aware
multi-plan routing (#57), model catalog overhaul, dashboard isDefault fix
- README/README_EN: replace outdated gpt-5.4 model table with current
11-model catalog (gpt-5.2-codex as new flagship), update Claude Code
setup section to remove stale model mappings, note gpt-5.4/gpt-5.3-codex
removal for free/plus accounts

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Files changed (3) hide show
  1. CHANGELOG.md +9 -0
  2. README.md +22 -19
  3. README_EN.md +22 -19
CHANGELOG.md CHANGED
@@ -9,9 +9,18 @@
9
  ### Added
10
 
11
  - 更新弹窗 + 自动重启:点击"有可用更新"弹出 Modal 显示 changelog,一键更新后服务器自动重启、前端自动刷新,零人工干预(git 模式 spawn 新进程、Docker/Electron 显示对应操作指引)
 
 
 
 
 
 
 
12
 
13
  ### Fixed
14
 
 
 
15
  - Docker 端口修复:锁定容器内 `PORT=8080`(`environment` 覆盖 `env_file`),HEALTHCHECK 固定检查 8080,`.env` 的 PORT 仅控制宿主机暴露端口,修复自定义 PORT 时健康检查失败和端口映射不匹配的问题 (#40)
16
  - Docker Compose 暴露 OAuth 回调端口 1455,修复容器内登录时 "Operation timed out" 的问题
17
  - README Docker 快速开始补充 `cp .env.example .env` 步骤,修复新用户因缺少 `.env` 文件导致 `docker compose up -d` 启动失败的问题 (#38)
 
9
  ### Added
10
 
11
  - 更新弹窗 + 自动重启:点击"有可用更新"弹出 Modal 显示 changelog,一键更新后服务器自动重启、前端自动刷新,零人工干预(git 模式 spawn 新进程、Docker/Electron 显示对应操作指引)
12
+ - Model-aware 多计划账号路由:不同 plan(free/plus/business)的账号自动路由到各自支持的模型,business 账号可继续使用 gpt-5.4 等高端模型 (#57)
13
+
14
+ ### Changed
15
+
16
+ - 模型目录大幅更新:后端移除 free/plus 账号的 `gpt-5.4`、`gpt-5.3-codex` 全系列,新旗舰模型为 `gpt-5.2-codex`(`codex` 别名指向此模型)
17
+ - 新增模型:`gpt-5.2`、`gpt-5.1-codex`、`gpt-5.1`、`gpt-5-codex`、`gpt-5`、`gpt-oss-120b`、`gpt-oss-20b`、`gpt-5-codex-mini`
18
+ - 模型目录从 23 个静态模型精简为 11 个(匹配后端实际返回)
19
 
20
  ### Fixed
21
 
22
+ - `cached_tokens` / `reasoning_tokens` 透传:从 Codex API 响应的 `input_tokens_details` 和 `output_tokens_details` 中提取,传递到 OpenAI(`prompt_tokens_details`)、Anthropic(`cache_read_input_tokens`)、Gemini(`cachedContentTokenCount`)三种格式,覆盖流式和非流式模式 (#55, #58)
23
+ - Dashboard 模型选择器使用后端 catalog 的 `isDefault` 字段,替代硬编码 `gpt-5.4`
24
  - Docker 端口修复:锁定容器内 `PORT=8080`(`environment` 覆盖 `env_file`),HEALTHCHECK 固定检查 8080,`.env` 的 PORT 仅控制宿主机暴露端口,修复自定义 PORT 时健康检查失败和端口映射不匹配的问题 (#40)
25
  - Docker Compose 暴露 OAuth 回调端口 1455,修复容器内登录时 "Operation timed out" 的问题
26
  - README Docker 快速开始补充 `cp .env.example .env` 步骤,修复新用户因缺少 `.env` 文件导致 `docker compose up -d` 启动失败的问题 (#38)
README.md CHANGED
@@ -169,17 +169,23 @@ curl http://localhost:8080/v1/chat/completions \
169
 
170
  | 模型 ID | 别名 | 推理等级 | 说明 |
171
  |---------|------|---------|------|
172
- | `gpt-5.4` | `codex` | minimal / low / medium / high | 最新旗舰编程模型(默认) |
173
- | `gpt-5.3-codex` | — | low / medium / high | 上一代旗舰 agentic 编程模型 |
174
- | `gpt-5.3-codex-spark` | — | minimal / low | 超轻量编程模型 |
175
- | `gpt-5.2-codex` | — | low / medium / high | agentic 编程模型 |
176
- | `gpt-5.1-codex-max` | `codex-max` | low / medium / high | 深度推理编程模型 |
177
- | `gpt-5.1-codex-mini` | `codex-mini` | low / medium / high | 轻量快速编程模型 |
 
 
 
 
 
178
 
179
  > **模型名后缀**:在任意模型名后追加 `-fast` 启用 Fast 模式,追加 `-high`/`-low` 等切换推理等级。
180
- > 例如:`gpt-5.4-fast`、`gpt-5.4-high-fast`、`codex-fast`
181
  >
182
- > 模型表会随 Codex Desktop 版本更新自动同步后端也会动态获取最新模型目录。
 
183
 
184
  ## 🔗 客户端接入 (Client Setup)
185
 
@@ -190,22 +196,19 @@ curl http://localhost:8080/v1/chat/completions \
190
  ```bash
191
  export ANTHROPIC_BASE_URL=http://localhost:8080
192
  export ANTHROPIC_API_KEY=your-api-key
193
- # 默认 Opus 4.6 → gpt-5.4,无需设置 ANTHROPIC_MODEL
194
  # 如需切换模型或启用后缀:
195
- # export ANTHROPIC_MODEL=codex-fast # → gpt-5.4 + Fast 模式
196
- # export ANTHROPIC_MODEL=gpt-5.4-high # → gpt-5.4 + high 推理
197
- # export ANTHROPIC_MODEL=gpt-5.4-high-fast # → gpt-5.4 + high + Fast
198
- # export ANTHROPIC_MODEL=claude-sonnet-4-6 # Sonnet gpt-5.3-codex
199
- # export ANTHROPIC_MODEL=claude-haiku-4-5-20251001 # Haiku → gpt-5.1-codex-mini
200
 
201
  claude # 启动 Claude Code
202
  ```
203
 
204
- | Claude Code 模型 | 映射到 Codex 模型 | 说明 |
205
- |------------------|------------------|------|
206
- | Opus (默认) | `gpt-5.4` | 无需设置 `ANTHROPIC_MODEL` |
207
- | Sonnet (`claude-sonnet-4-6`) | `gpt-5.3-codex` | |
208
- | Haiku (`claude-haiku-4-5-20251001`) | `gpt-5.1-codex-mini` | |
209
 
210
  > 也可以在控制面板 (`http://localhost:8080`) 的 **Anthropic SDK Setup** 卡片中一键复制环境变量。
211
 
 
169
 
170
  | 模型 ID | 别名 | 推理等级 | 说明 |
171
  |---------|------|---------|------|
172
+ | `gpt-5.2-codex` | `codex` | low / medium / high / xhigh | 前沿 agentic 编程模型(默认) |
173
+ | `gpt-5.2` | — | low / medium / high / xhigh | 专业工作 + 长时间代理 |
174
+ | `gpt-5.1-codex-max` | — | low / medium / high / xhigh | 扩展上下文 / 深度推理 |
175
+ | `gpt-5.1-codex` | — | low / medium / high | GPT-5.1 编程模型 |
176
+ | `gpt-5.1` | | low / medium / high | 通用 GPT-5.1 |
177
+ | `gpt-5-codex` | | low / medium / high | GPT-5 编程模型 |
178
+ | `gpt-5` | — | minimal / low / medium / high | 通用 GPT-5 |
179
+ | `gpt-oss-120b` | — | low / medium / high | 开源 120B 模型 |
180
+ | `gpt-oss-20b` | — | low / medium / high | 开源 20B 模型 |
181
+ | `gpt-5.1-codex-mini` | — | medium / high | 轻量快速编程模型 |
182
+ | `gpt-5-codex-mini` | — | medium / high | 轻量编程模型 |
183
 
184
  > **模型名后缀**:在任意模型名后追加 `-fast` 启用 Fast 模式,追加 `-high`/`-low` 等切换推理等级。
185
+ > 例如:`codex-fast`、`gpt-5.2-codex-high-fast`。
186
  >
187
+ > **注意**:`gpt-5.4`、`gpt-5.3-codex` 系已从 free/plus 账号移除,仅 business 账号可用
188
+ > 模型列表由后端动态获取,会自动同步最新可用模型。
189
 
190
  ## 🔗 客户端接入 (Client Setup)
191
 
 
196
  ```bash
197
  export ANTHROPIC_BASE_URL=http://localhost:8080
198
  export ANTHROPIC_API_KEY=your-api-key
199
+ # 默认使用 gpt-5.2-codex(codex 别名),无需设置 ANTHROPIC_MODEL
200
  # 如需切换模型或启用后缀:
201
+ # export ANTHROPIC_MODEL=codex-fast # → gpt-5.2-codex + Fast 模式
202
+ # export ANTHROPIC_MODEL=codex-high # → gpt-5.2-codex + high 推理
203
+ # export ANTHROPIC_MODEL=codex-high-fast # → gpt-5.2-codex + high + Fast
204
+ # export ANTHROPIC_MODEL=gpt-5.2 # → 通用 GPT-5.2
205
+ # export ANTHROPIC_MODEL=gpt-5.1-codex-mini # → 轻量快速模型
206
 
207
  claude # 启动 Claude Code
208
  ```
209
 
210
+ > 所有 Claude Code 模型名(Opus / Sonnet / Haiku)均映射到配置的默认模型(`gpt-5.2-codex`)。
211
+ > 如需指定具体模型,通过 `ANTHROPIC_MODEL` 环境变量设置 Codex 模型名即可。
 
 
 
212
 
213
  > 也可以在控制面板 (`http://localhost:8080`) 的 **Anthropic SDK Setup** 卡片中一键复制环境变量。
214
 
README_EN.md CHANGED
@@ -169,17 +169,23 @@ curl http://localhost:8080/v1/chat/completions \
169
 
170
  | Model ID | Alias | Reasoning Efforts | Description |
171
  |----------|-------|-------------------|-------------|
172
- | `gpt-5.4` | `codex` | minimal / low / medium / high | Latest flagship coding model (default) |
173
- | `gpt-5.3-codex` | — | low / medium / high | Previous-gen flagship agentic coding model |
174
- | `gpt-5.3-codex-spark` | — | minimal / low | Ultra-lightweight coding model |
175
- | `gpt-5.2-codex` | — | low / medium / high | Agentic coding model |
176
- | `gpt-5.1-codex-max` | `codex-max` | low / medium / high | Deep reasoning coding model |
177
- | `gpt-5.1-codex-mini` | `codex-mini` | low / medium / high | Lightweight, fast coding model |
 
 
 
 
 
178
 
179
  > **Model name suffixes**: Append `-fast` to any model name to enable Fast mode, or `-high`/`-low` etc. to change reasoning effort.
180
- > Examples: `gpt-5.4-fast`, `gpt-5.4-high-fast`, `codex-fast`.
181
  >
182
- > Models are automatically synced when new Codex Desktop versions are released. The backend also dynamically fetches the latest model catalog.
 
183
 
184
  ## 🔗 Client Setup
185
 
@@ -190,22 +196,19 @@ Set environment variables to route Claude Code through codex-proxy:
190
  ```bash
191
  export ANTHROPIC_BASE_URL=http://localhost:8080
192
  export ANTHROPIC_API_KEY=your-api-key
193
- # Default Opus 4.6 gpt-5.4, no need to set ANTHROPIC_MODEL
194
  # To switch models or use suffixes:
195
- # export ANTHROPIC_MODEL=codex-fast # → gpt-5.4 + Fast mode
196
- # export ANTHROPIC_MODEL=gpt-5.4-high # → gpt-5.4 + high reasoning
197
- # export ANTHROPIC_MODEL=gpt-5.4-high-fast # → gpt-5.4 + high + Fast
198
- # export ANTHROPIC_MODEL=claude-sonnet-4-6 # Sonnet gpt-5.3-codex
199
- # export ANTHROPIC_MODEL=claude-haiku-4-5-20251001 # Haiku → gpt-5.1-codex-mini
200
 
201
  claude # Launch Claude Code
202
  ```
203
 
204
- | Claude Code Model | Maps to Codex Model | Notes |
205
- |-------------------|---------------------|-------|
206
- | Opus (default) | `gpt-5.4` | No need to set `ANTHROPIC_MODEL` |
207
- | Sonnet (`claude-sonnet-4-6`) | `gpt-5.3-codex` | |
208
- | Haiku (`claude-haiku-4-5-20251001`) | `gpt-5.1-codex-mini` | |
209
 
210
  > You can also copy environment variables from the **Anthropic SDK Setup** card in the dashboard (`http://localhost:8080`).
211
 
 
169
 
170
  | Model ID | Alias | Reasoning Efforts | Description |
171
  |----------|-------|-------------------|-------------|
172
+ | `gpt-5.2-codex` | `codex` | low / medium / high / xhigh | Frontier agentic coding model (default) |
173
+ | `gpt-5.2` | — | low / medium / high / xhigh | Professional work & long-running agents |
174
+ | `gpt-5.1-codex-max` | — | low / medium / high / xhigh | Extended context / deepest reasoning |
175
+ | `gpt-5.1-codex` | — | low / medium / high | GPT-5.1 coding model |
176
+ | `gpt-5.1` | | low / medium / high | General-purpose GPT-5.1 |
177
+ | `gpt-5-codex` | | low / medium / high | GPT-5 coding model |
178
+ | `gpt-5` | — | minimal / low / medium / high | General-purpose GPT-5 |
179
+ | `gpt-oss-120b` | — | low / medium / high | Open-source 120B model |
180
+ | `gpt-oss-20b` | — | low / medium / high | Open-source 20B model |
181
+ | `gpt-5.1-codex-mini` | — | medium / high | Lightweight, fast coding model |
182
+ | `gpt-5-codex-mini` | — | medium / high | Lightweight coding model |
183
 
184
  > **Model name suffixes**: Append `-fast` to any model name to enable Fast mode, or `-high`/`-low` etc. to change reasoning effort.
185
+ > Examples: `codex-fast`, `gpt-5.2-codex-high-fast`.
186
  >
187
+ > **Note**: `gpt-5.4` and `gpt-5.3-codex` families have been removed for free/plus accounts. Only business accounts retain access.
188
+ > Models are dynamically fetched from the backend and will automatically sync the latest available catalog.
189
 
190
  ## 🔗 Client Setup
191
 
 
196
  ```bash
197
  export ANTHROPIC_BASE_URL=http://localhost:8080
198
  export ANTHROPIC_API_KEY=your-api-key
199
+ # Default model is gpt-5.2-codex (codex alias), no need to set ANTHROPIC_MODEL
200
  # To switch models or use suffixes:
201
+ # export ANTHROPIC_MODEL=codex-fast # → gpt-5.2-codex + Fast mode
202
+ # export ANTHROPIC_MODEL=codex-high # → gpt-5.2-codex + high reasoning
203
+ # export ANTHROPIC_MODEL=codex-high-fast # → gpt-5.2-codex + high + Fast
204
+ # export ANTHROPIC_MODEL=gpt-5.2 # → General-purpose GPT-5.2
205
+ # export ANTHROPIC_MODEL=gpt-5.1-codex-mini # → Lightweight, fast model
206
 
207
  claude # Launch Claude Code
208
  ```
209
 
210
+ > All Claude Code model names (Opus / Sonnet / Haiku) map to the configured default model (`gpt-5.2-codex`).
211
+ > To use a specific model, set the `ANTHROPIC_MODEL` environment variable to a Codex model name.
 
 
 
212
 
213
  > You can also copy environment variables from the **Anthropic SDK Setup** card in the dashboard (`http://localhost:8080`).
214