[BOT ISSUE] Groq: add deprecation markers for 3 models scheduled for shutdown

[github-actions]

Summary

Three Groq models have officially announced shutdown dates but lack deprecation_date markers in the catalog. Without the deprecation_date field, markModelsPastDeprecationDate() in packages/proxy/schema/models.ts will not auto-flag these as deprecated, and they will continue appearing as active models in the UI past their shutdown dates.

Two of these models (qwen/qwen3-32b and meta-llama/llama-4-scout-17b-16e-instruct) are shutting down on July 17, 2026 — only 16 days from today.

Affected Models

Model ID	Current deprecation_date	Correct Value	Shutdown Date	Replacement	Line
qwen/qwen3-32b	(missing)	2026-07-17	07/17/2026	openai/gpt-oss-120b	~3244
meta-llama/llama-4-scout-17b-16e-instruct	(missing)	2026-07-17	07/17/2026	openai/gpt-oss-120b or qwen/qwen3.6-27b	~4720
llama-3.3-70b-versatile	(missing)	2026-08-16	08/16/2026	openai/gpt-oss-120b or qwen/qwen3.6-27b	~4663

Also affected (same deprecation schedule):

• llama-3.1-8b-instant (line ~4675) — shutdown 08/16/2026, replacement: openai/gpt-oss-20b

All other fields (pricing, format, token limits) are correct and unchanged. Only the deprecation marker is missing.

Verification

Check	Status	Detail
Cross-source (deprecation dates)	✅	Groq models page: lists models as "Preview" with lifecycle info AND Groq deprecations page: explicit shutdown dates and replacement models for all 4
Already fixed	✅	Confirmed all 4 entries lack deprecation_date in model_list.json
Replacement models exist	✅	openai/gpt-oss-120b, openai/gpt-oss-20b, and qwen/qwen3.6-27b are all present in the catalog
Duplicate check	✅	No open issue covers Groq deprecation markers

Verification Notes

Field	Source	Notes
qwen/qwen3-32b shutdown 07/17/2026	Groq deprecations	Replacement: openai/gpt-oss-120b
llama-4-scout shutdown 07/17/2026	Groq deprecations	Replacement: openai/gpt-oss-120b or qwen/qwen3.6-27b
llama-3.3-70b-versatile shutdown 08/16/2026	Groq deprecations	Replacement: openai/gpt-oss-120b or qwen/qwen3.6-27b
llama-3.1-8b-instant shutdown 08/16/2026	Groq deprecations	Replacement: openai/gpt-oss-20b

Local Files Inspected

• packages/proxy/schema/model_list.json — confirmed all 4 entries lack deprecation_date
• packages/proxy/schema/models.ts — markModelsPastDeprecationDate() (line ~222) checks deprecation_date per entry

{
  "kind": "stale_metadata",
  "provider": "groq",
  "models": ["qwen/qwen3-32b", "meta-llama/llama-4-scout-17b-16e-instruct", "llama-3.3-70b-versatile"],
  "status": "deprecated",
  "model_specs": {
    "qwen/qwen3-32b": {
      "format": "openai",
      "flavor": "chat",
      "input_cost_per_mil_tokens": 0.29,
      "output_cost_per_mil_tokens": 0.59,
      "displayName": "Qwen3-32B",
      "reasoning": true,
      "experimental": true,
      "deprecation_date": "2026-07-17",
      "max_input_tokens": 131072,
      "max_output_tokens": 40960,
      "available_providers": ["groq"]
    },
    "meta-llama/llama-4-scout-17b-16e-instruct": {
      "format": "openai",
      "flavor": "chat",
      "input_cost_per_mil_tokens": 0.11,
      "output_cost_per_mil_tokens": 0.34,
      "displayName": "Llama 4 Scout (17Bx16E)",
      "experimental": true,
      "deprecation_date": "2026-07-17",
      "max_input_tokens": 131072,
      "max_output_tokens": 8192,
      "available_providers": ["groq"]
    },
    "llama-3.3-70b-versatile": {
      "format": "openai",
      "flavor": "chat",
      "input_cost_per_mil_tokens": 0.59,
      "output_cost_per_mil_tokens": 0.79,
      "displayName": "Llama 3.3 70B Versatile 128k",
      "deprecation_date": "2026-08-16",
      "max_input_tokens": 128000,
      "max_output_tokens": 32768,
      "available_providers": ["groq"]
    }
  },
  "source_urls": [
    "https://console.groq.com/docs/models",
    "https://console.groq.com/docs/deprecations"
  ]
}

[github-actions]

Prepared catalog updates for qwen/qwen3-32b, meta-llama/llama-4-scout-17b-16e-instruct, llama-3.3-70b-versatile. Updated models: qwen/qwen3-32b, meta-llama/llama-4-scout-17b-16e-instruct, llama-3.3-70b-versatile.

Verification sources: https://console.groq.com/docs/models, https://console.groq.com/docs/deprecations
Verified metadata applied:

• qwen/qwen3-32b: format=openai, flavor=chat, displayName=Qwen3-32B, max_input_tokens=131072, max_output_tokens=40960, pricing(input/output per 1M)=0.29/0.59, providers=groq, reasoning=true, deprecation_date=2026-07-17
• meta-llama/llama-4-scout-17b-16e-instruct: format=openai, flavor=chat, displayName=Llama 4 Scout (17Bx16E), max_input_tokens=131072, max_output_tokens=8192, pricing(input/output per 1M)=0.11/0.34, providers=groq, deprecation_date=2026-07-17
• llama-3.3-70b-versatile: format=openai, flavor=chat, displayName=Llama 3.3 70B Versatile 128k, max_input_tokens=128000, max_output_tokens=32768, pricing(input/output per 1M)=0.59/0.79, providers=groq, deprecation_date=2026-08-16

Included in the daily batch PR: #904