Model list for Jiekou.ai Outdated and Unsynced

[FallenerFurtherstr]

Jiekou.ai has gone through significant update on their model list. Namely, the recent GPT-5.6 models are available, among with others.
I saw the list for Jiekou on models.dev isn't updated at all, with 61 models (now 176).
So, can this be updated somehow?
And I saw a CI workflow called sync-models, but it doesn't have Jiekou.ai on its list.
It would be great if we have it on the sync list.
Attached is the model list json that I curled from Jiekou.ai via https://api.jiekou.ai/openai/v1/models:

JiekouAIModelList260713.json

[github-actions]

Now I have a complete picture of the situation. Let me summarize my analysis and decision.

Analysis

The issue has two distinct requests:

1. Update the Jiekou model list — The catalog has 61 model TOMLs, but Jiekou's API now exposes ~144 models. After applying the generator script's SKIP_MODELS patterns, 52 non-skipped models are missing from the catalog (across Claude, GPT-5.x, Gemini 3.x, Qwen 3.5, GLM-5, DeepSeek V4, MiniMax M2.5/M2.7, and Xiaomi MiMo families).

2. Add Jiekou.ai to the sync-models CI workflow — Jiekou currently has only a standalone generator script (providers/jiekou/scripts/generate-models.ts) and is NOT registered in the centralized sync system (packages/core/src/sync/index.ts).

Decision

This issue needs maintainer review — no automated fix was opened. Here's why:

The CI integration request is a code/infrastructure change, not a catalog data fix. Adding Jiekou to sync-models requires creating a new sync provider module (packages/core/src/sync/providers/jiekou.ts), registering it in packages/core/src/sync/index.ts, and potentially adding secrets to .github/workflows/sync-models.yml. My instructions explicitly exclude code, schema, and workflow changes.

The bulk model update is an operational task, not a minimal TOML correction. The 52 missing models should be generated by running the existing providers/jiekou/scripts/generate-models.ts script (which I cannot do without Bash) or by integrating Jiekou into the centralized sync system. Hand-creating 52 TOML files would not be a minimal change and would duplicate the generator's purpose.

Data quality concerns prevent a safe manual fix:

• The existing hand-curated Jiekou TOMLs include fields the API does not provide: human-written description, model-specific reasoning_options (e.g., [{ type = "effort", values = [...] }, { type = "budget_tokens", ... }]), knowledge cutoffs, and accurate release_date/family values. The generator script omits all of these, and AGENTS.md requires that reasoning models declare reasoning_options.
• Many missing model IDs appear to be Jiekou-specific gateway aliases (e.g., gpt-5.6-terra/luna/sol, claude-opus-4-8-r/-cc/-dd, claude-fable-5) that cannot be cross-referenced against first-party OpenAI/Anthropic documentation. Properly cataloging each requires understanding what underlying model each alias maps to and what reasoning controls Jiekou exposes for it.
• The API response (~190 KB) was truncated when fetched; the reconstructed model list has methodological caveats and may not be fully accurate for all 52 entries.

Recommended path for the maintainer: Integrate Jiekou into the centralized sync system (following sync.md's "Adding A Provider" steps), which would give it hourly automated sync like the other 21 providers. This would also require enhancing the sync module to handle Jiekou's tiered billing (is_tiered_billing/tiered_billing fields seen in the API) and to emit reasoning_options consistent with the provider documentation already captured in providers/jiekou/provider.toml.