fix(channel-test): auto-detect image generation models

[lichengzhe]

⚠️ 提交说明 / PR Notice

Important

• 请提供人工撰写的简洁摘要，避免直接粘贴未经整理的 AI 输出。

📝 变更描述 / Description

渠道测试使用「自动检测」时，gpt-image-2 会被当成聊天模型，请求发到 /v1/chat/completions，导致上游报错（#6121）。

本 PR 做了这些调整：

• 将 gpt-image-1 的精确匹配改为匹配整个 gpt-image-* 系列，同时支持 openai/gpt-image-3 这类带命名空间的名称。
• 渠道测试统一使用模型分类器。识别为图像模型后，会同时使用 /v1/images/generations 和 dto.ImageRequest，避免请求路径与请求体不一致。
• 手动指定的端点仍然优先。
• 保留火山引擎 Seedream 的自动识别，并改为不区分大小写，Seedream-4.0 也能正常识别。

自动检测会对所有渠道类型生效。dall-e、flux、imagen 等已知图像模型也会自动走图像端点。如果个别上游通过聊天接口代理图像模型，仍可手动指定端点覆盖。

定价页复用同一个分类器，因此 gpt-image-2 的端点类型也会显示为 [image-generation, openai]。

本 PR 在 AI 辅助下完成，实现、测试、影响范围与本描述均经本人复核。

🚀 变更类型 / Type of change

• 🐛 Bug 修复 (Bug fix) - 请关联对应 Issue，避免将设计取舍、理解偏差或预期不一致直接归类为 bug
• ✨ 新功能 (New feature) - 重大特性建议先通过 Issue 沟通
• ⚡ 性能优化 / 重构 (Refactor)
• 📝 文档更新 (Documentation)

🔗 关联任务 / Related Issue

• Closes bug: channel test auto-detection sends gpt-image-2 to chat completions #6121

✅ 提交前检查项 / Checklist

• 人工确认: 我已亲自整理并撰写此描述，没有直接粘贴未经处理的 AI 输出。
• 非重复提交: 我已搜索现有的 Issues 与 PRs，确认不是重复提交。
• Bug fix 说明: 若此 PR 标记为 Bug fix，我已提交或关联对应 Issue，且不会将设计取舍、预期不一致或理解偏差直接归类为 bug。
• 变更理解: 我已理解这些更改的工作原理及可能影响。
• 范围聚焦: 本 PR 未包含任何与当前任务无关的代码改动。
• 本地验证: 已在本地运行并通过测试或手动验证，维护者可以据此复核结果。
• 安全合规: 代码中无敏感凭据，且符合项目代码规范。

📸 运行证明 / Proof of Work

以下测试已通过：

go test ./common ./controller -count=1
ok github.com/QuantumNous/new-api/common
ok github.com/QuantumNous/new-api/controller

回归测试覆盖：

• gpt-image-1、gpt-image-2、带命名空间及未来的 gpt-image-* 名称
• 手动指定端点的优先级
• 图像请求路径与 dto.ImageRequest 保持一致
• 火山引擎 Seedream，包括 Seedream-4.0 的大小写场景
• 普通聊天模型行为不变
• 定价元数据正确返回 [image-generation, openai]

其余 Go 包的测试也已通过。只有仓库根目录的 setup 因本地缺少生成产物 web/classic/dist 无法运行，与本次改动无关。

Summary by CodeRabbit

• New Features

  • Added broader support for GPT image model variants, including newer and differently formatted model names.
  • Image-generation requests are now automatically detected and routed to the appropriate image endpoint.
  • Image model testing now uses the expected image request format and default settings.

• Bug Fixes

  • Improved endpoint selection for image-generation models across supported providers, while preserving automatic handling for chat models.

[coderabbitai]

Walkthrough

GPT image model detection now covers the gpt-image- family. Endpoint preference and channel-test auto-detection route these models to image-generation requests while preserving explicit endpoint selection.

Changes

Image endpoint detection

Layer / File(s)	Summary
GPT image model classification and endpoint preference common/model.go, common/model_test.go	GPT image model matching uses the gpt-image- family prefix, with tests covering model recognition and image-generation endpoint ordering.
Channel-test image request routing controller/channel-test.go, controller/channel_test_internal_test.go	Automatic channel-test detection maps image models to the image-generation endpoint, removes the Seedream-specific path override, and validates image probe construction and explicit endpoint precedence.

Estimated code review effort: 3 (Moderate) | ~20 minutes

Possibly related PRs

• QuantumNous/new-api#1814: Also changes channel-test routing for image-generation endpoints and Seedream handling.

Poem

A rabbit hops past chat’s old door,
“GPT images belong by /images/generations floor!”
Prefixes bloom, probes shape up,
Explicit choices stay in the cup.
Carrots for tests—hip-hop hooray!

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Linked Issues check	✅ Passed	The changes address issue #6121 by classifying gpt-image-* as image-generation, routing auto-detect to images, and preserving explicit endpoint precedence.
Out of Scope Changes check	✅ Passed	The modified code and tests stay focused on image-generation auto-detection and related pricing metadata, with no unrelated changes evident.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title is concise and accurately summarizes the main change to auto-detect image generation models.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}