fix(agent): allow CherryIN OpenAI-protocol models in Agent model picker

[Azure12355]

What this PR does

Before this PR:
In the Agent model picker, CherryIN's openai/gpt-* models were silently excluded from search results even though they are enabled in Provider settings and work correctly in the regular Chat picker.

After this PR:
CherryIN's OpenAI-protocol models (e.g. openai/gpt-5.x) now appear in the Agent model picker alongside Anthropic-protocol models, matching Chat picker behavior.

Fixes #14717

Why we need it and why it was done in this way

getProviderAnthropicModelChecker previously grouped cherryin with new-api, restricting both to endpoint_type === 'anthropic'. However, CherryIN is a first-party aggregator that serves both Anthropic and OpenAI models — the same role as aihubmix, which already uses return () => true.

The fix separates cherryin into its own case with return () => true, matching aihubmix behavior. new-api retains its strict anthropic-only filter because it is a generic OpenAI-compatible gateway without the same first-party guarantees.

The following tradeoffs were made:

• Only cherryin is relaxed; new-api is intentionally kept strict to avoid unintended side-effects on other deployments.

The following alternatives were considered:

• Relaxing new-api as well — rejected, too broad and potentially breaking for other new-api deployments.

Breaking changes

None.

Special notes for your reviewer

The change is a single case split in getProviderAnthropicModelChecker (src/main/apiServer/utils/index.ts). No schema, UI, or IPC changes.

Checklist

• PR: The PR description is expressive enough and will help future contributors
• Code: Write code that humans can understand and Keep it simple
• Refactor: You have left the code cleaner than you found it (Boy Scout Rule)
• Upgrade: Impact of this change on upgrade flows was considered and addressed if required
• Documentation: A user-guide update was considered and is present (link) or not required. Check this only when the PR introduces or changes a user-facing feature or behavior.
• Self-review: I have reviewed my own code before requesting review from others

Release note

Fixed: CherryIN's OpenAI-protocol models (openai/gpt-*) now appear in the Agent model picker, matching their visibility in the regular Chat picker.

[DeJeune]

Note

This comment was translated by Claude.

Have you tested it?

---

Original Content

测试过吗

[Azure12355]

Have you tested it?

Note

This comment was translated by Claude.

Yes, I have tested it.

---

Original Content

Have you tested it?

测试过了

[zhibisora]

Review summary

Thanks for the focused fix. I think this needs one change before merge: the current CherryIN branch is too broad for an Anthropic-only Agent model picker.

Requested changes

Please replace the unconditional CherryIN allow-list with a shared Anthropic endpoint capability check for both cherryin and new-api, using supported_endpoint_types?.includes('anthropic') plus endpoint_type === 'anthropic' as a compatibility fallback for manually added or older models.

[zhibisora]

LGTM. Verified the follow-up commit addresses the Anthropic endpoint filtering concern by using supported_endpoint_types with the legacy endpoint_type fallback. CI checks are passing.