chore: remove deprecated models from the catalog

[github-actions]

Model deprecation audit

Each model below returned a definitive not-found / deprecated response from its provider (a single 404 / not-found is treated as authoritative). Account/region-scoped providers (bedrock, vertex, databricks) are surfaced for manual review, not auto-removed.

{
  "deprecations": [
    {
      "model": "gpt-5.6",
      "provider": "azure",
      "reason": "absent from list + probe not-found",
      "status": 400,
      "detail": "{\n  \"error\": {\n    \"message\": \"Unsupported parameter: 'max_tokens' is not supported with this model. Use 'max_completion_tokens' instead.\",\n    \"type\": \"invalid"
    },
    {
      "model": "gpt-5.6",
      "provider": "openai",
      "reason": "absent from list + probe not-found",
      "status": 400,
      "detail": "{\n  \"error\": {\n    \"message\": \"Unsupported parameter: 'max_tokens' is not supported with this model. Use 'max_completion_tokens' instead.\",\n    \"type\": \"invalid"
    }
  ],
  "reportOnly": [
    {
      "provider": "bedrock",
      "modelCount": 117
    },
    {
      "provider": "databricks",
      "modelCount": 36
    },
    {
      "provider": "fireworks",
      "modelCount": 265
    },
    {
      "provider": "vertex",
      "modelCount": 117
    }
  ],
  "skipped": [
    {
      "provider": "lepton",
      "reason": "no adapter"
    },
    {
      "provider": "replicate",
      "reason": "no adapter"
    }
  ]
}

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
ai-proxy	Ready	Preview, Comment	Jul 11, 2026 7:03am

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 657441c873

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "Codex (@codex) review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "Codex (@codex) address that feedback".

[chatgpt-codex-connector]

Do not exclude gpt-5.6 on a parameter error

The audit report in this commit shows the gpt-5.6 probes returned 400 Unsupported parameter: 'max_tokens' ... Use 'max_completion_tokens' instead, not a provider not-found/deprecated response. Since sync_models.ts folds this JSON into SYNC_EXCLUDED_MODELS and the same commit removes the catalog entry, leaving this line blocks the model from being restored by sync; fetchOpenAI already rewrites max_tokens for GPT-5 reasoning models before forwarding, so users selecting gpt-5.6 can lose model discovery//auto access because the probe used the wrong parameter.

Useful? React with 👍 / 👎.