Skip to content

Commit dfee892

Browse files
committed
docs(changelog): 新增 v0.5.0 版本日志与 Model Calling 功能截图;
🤖 Generated with [Claude Code](https://github.com/claude), [CodeX](https://openai.com), [Gemini](https://github.com/apps/gemini-code-assist) Co-Authored-By: Aurelius Huang<threefish.ai@gmail.com>
1 parent c6665da commit dfee892

2 files changed

Lines changed: 23 additions & 1 deletion

File tree

CHANGELOG.md

Lines changed: 23 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -4,7 +4,29 @@
44

55
## [Unreleased]
66

7-
- feat(zhipu): 新增每模型并发限制(默认 3,可通过 `vendors[zhipu].concurrency` 配置),基于 `asyncio.Semaphore` 实现 FIFO 公平排队,流式与非流式共用同一槽位,与 429 重试机制兼容。
7+
## [v0.5.0](https://github.com/ThreeFish-AI/coding-proxy/releases/tag/v0.5.0) - 2026-05-27
8+
9+
> [!IMPORTANT]
10+
>
11+
> **🚀 Model Calling 实时状态!**
12+
>
13+
> 模型并发与排队深度一目了然,运行时动态调整每个模型并行度,预防 vendor 侧的 429 幺蛾子。
14+
15+
![model-calling](assets/model-calling-v0.5.0.png)
16+
17+
### ✨ 核心亮点
18+
19+
- feat(concurrency): 新增 Model Calling 实时状态模块,可视化每模型并发与排队深度,支持运行时动态修改每模型并行度 (#250) (#251)
20+
- feat(zhipu): 新增每模型并发限制,默认 3 个并行请求 FIFO 排队 (#248)
21+
- feat(zhipu): 为 429 Rate Limit 添加指数退避重试挽回机制 (#242)
22+
23+
### 🔧 更多特性
24+
25+
- fix(antigravity): 修复 v1internal 模式检测逻辑并新增 E2E 测试; (#234)
26+
- fix(routes): 修复 count_tokens 路由对 target_vendor.name 的错误属性访问; (#235)
27+
- fix(vendor-channels): 修复 zhipu→anthropic 通道 tool_use/tool_result 配对漏洞; (#236)
28+
- fix(native-api): 修复 Gemini :verb 路径中 %3A URL 编码导致上游 400 的兼容问题; (#237)
29+
- fix(zhipu): 诊断首选 tier 语义拒绝降级问题,增强可观测性并提取跨供应商清洗共享函数 (#243)
830

931
## [v0.4.0](https://github.com/ThreeFish-AI/coding-proxy/releases/tag/v0.4.0) — 2026-05-01
1032

assets/model-calling-v0.5.0.png

21.9 KB
Loading

0 commit comments

Comments
 (0)