From 78728e61ed06b5e7b2fd623c61d13d1811200d6a Mon Sep 17 00:00:00 2001
From: Yige <a@wyr.me>
Date: Wed, 13 May 2026 18:54:42 +0800
Subject: [PATCH] =?UTF-8?q?fix(i18n):=20GPT-5=20/=20o-series=20=E4=B8=8D?=
 =?UTF-8?q?=E6=94=AF=E6=8C=81=E8=87=AA=E5=AE=9A=E4=B9=89=20temperature=20(?=
 =?UTF-8?q?#11)?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

## 接 PR #8 的进一步根因

本地用 personal PAT 端到端验证时发现：付费 quota 已生效、`gpt-5-mini`
可访问、`max_completion_tokens` 修复正确，但 translate.py 仍 400：
```
"message": "Unsupported value: 'temperature' does not support 0.1 with this model. Only the default (1) value is supported."
"param": "temperature"
"code": "unsupported_value"
```

GPT-5 系列和 o-series（o1 / o3 / o4）作为 reasoning models 强制
`temperature=1`，不允许自定义。

## 改动

`call_github_models` 基于 model id 条件设置 `temperature`：
- `gpt-5*` / `*o1*` / `*o3*` / `*o4*` → 不发该字段（用默认 1）
- 其他 model（gpt-4 系列、Anthropic Claude）→ 保持 `temperature=0.1`（翻译任务偏好
deterministic）

## 验证

本地实际跑通：`GITHUB_TOKEN=$(gh auth token) uv run --quiet
scripts/i18n/translate.py skills/markdown`
```
skills/markdown → en-US:
  - calling github/openai/gpt-5-mini for en-US translation ...
  - wrote root SKILL.md with translated body (3838 chars)
```

gpt-5-mini 返回 HTTP 200，3838 字符英文 body 正确写入，Markdown 结构保留。
---
 scripts/i18n/translate.py | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/scripts/i18n/translate.py b/scripts/i18n/translate.py
index a5837af..c2f8fd9 100755
--- a/scripts/i18n/translate.py
+++ b/scripts/i18n/translate.py
@@ -191,15 +191,22 @@ def call_github_models(system_prompt: str, user_payload: str, model: str, endpoi
     # GPT-5 series rejects the legacy `max_tokens` field and requires
     # `max_completion_tokens` instead (OpenAI Chat Completions 2024+ contract).
     # GPT-4 and earlier accept either, so always use the new name.
-    payload = {
+    payload: dict[str, Any] = {
         "model": model,
         "messages": [
             {"role": "system", "content": system_prompt},
             {"role": "user", "content": user_payload},
         ],
-        "temperature": 0.1,
         "max_completion_tokens": 8192,
     }
+    # GPT-5 / o-series only allow the default temperature (1); sending any other
+    # value returns 400 `unsupported_value`. Older models accept custom values,
+    # where 0.1 is preferred for deterministic translation output.
+    is_gpt5_or_o_series = any(
+        tag in model.lower() for tag in ("gpt-5", "/o1", "/o3", "/o4")
+    )
+    if not is_gpt5_or_o_series:
+        payload["temperature"] = 0.1
     headers = {
         "Authorization": f"Bearer {token}",
         "Content-Type": "application/json",