[BOT ISSUE] Vertex: re-add missing Llama 4 Maverick and Scout MaaS entries (regression from #980)

[github-actions]

Provider

Vertex AI

Description

The Vertex AI pricing page lists publishers/meta/models/llama-4-maverick-17b-128e-instruct-maas and publishers/meta/models/llama-4-scout-17b-16e-instruct-maas as active MaaS models with per-token pricing, but neither exists in model_list.json.

These models were previously added via issue #980 / PR #1026 (closed 2026-07-27) but appear to have regressed — possibly removed by a subsequent LiteLLM sync.

Other Llama 4 entries exist under different provider formats (meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8 for Together, databricks-llama-4-maverick for Databricks), but the Vertex MaaS entries are absent.

Missing Models

Model ID	Input $/1M	Output $/1M	Multimodal	Source
publishers/meta/models/llama-4-maverick-17b-128e-instruct-maas	$0.50	$1.50	Yes	Vertex AI Pricing
publishers/meta/models/llama-4-scout-17b-16e-instruct-maas	$0.20	$0.60	Yes	Vertex AI Pricing

Suggested Catalog Entries

"publishers/meta/models/llama-4-maverick-17b-128e-instruct-maas": {
  "format": "openai",
  "flavor": "chat",
  "multimodal": true,
  "input_cost_per_mil_tokens": 0.5,
  "output_cost_per_mil_tokens": 1.5,
  "displayName": "Llama 4 Maverick (17Bx128E) Instruct",
  "max_input_tokens": 524288,
  "max_output_tokens": 8192,
  "available_providers": ["vertex"],
  "supported_regions": ["us-east5"]
}

"publishers/meta/models/llama-4-scout-17b-16e-instruct-maas": {
  "format": "openai",
  "flavor": "chat",
  "multimodal": true,
  "input_cost_per_mil_tokens": 0.2,
  "output_cost_per_mil_tokens": 0.6,
  "displayName": "Llama 4 Scout (17Bx16E) Instruct",
  "max_input_tokens": 1310720,
  "max_output_tokens": 8192,
  "available_providers": ["vertex"],
  "supported_regions": ["us-east5"]
}

Context window and max output values from the previous fix attempt (PR #1026 comment).

Verification

• Cross-source: model IDs and pricing confirmed on Vertex AI Pricing page; previously verified via Partner Models docs in issue [BOT ISSUE] Vertex: add missing Meta Llama 4 Maverick and Scout MaaS entries #980
• Already-fixed check: grep 'llama-4.*maas' model_list.json returns 0 matches — entries do not exist despite [BOT ISSUE] Vertex: add missing Meta Llama 4 Maverick and Scout MaaS entries #980 being closed
• ID format: matches existing publishers/meta/models/llama-3.3-70b-instruct-maas convention

Note

Issue #980 was closed as completed on 2026-07-27 after PR #1026 added these entries. They appear to have been removed by a subsequent sync. Consider adding these model IDs to SYNC_PRESERVED_FIELDS in sync_models.ts to prevent future regressions.

[github-actions]

Closing this bot issue because publishers/meta/models/llama-4-maverick-17b-128e-instruct-maas, publishers/meta/models/llama-4-scout-17b-16e-instruct-maas are on the permanent sync-exclusion list (MANUAL_SYNC_EXCLUDED_MODELS / deprecated_model_ids.json in sync_models.ts) or a Perplexity Gateway (router) model the proxy cannot route, and must not be auto-added to the catalog.