[BOT ISSUE] Databricks: add missing `databricks-gpt-5-nano`, `databricks-gpt-5-3-codex`, `databricks-gpt-5-2-codex`

[github-actions]

Summary

Three Databricks Foundation Model API pay-per-token GPT models are missing from packages/proxy/schema/model_list.json. This is a follow-up to #559, #585, #591, #592, #638, #639, and #640, continuing to address the Databricks catalog gap.

Missing Model IDs

Model ID	Context Window	Max Output	Supported Inputs	Description
databricks-gpt-5-nano	400,000	128,000	text, image	Excels at high-throughput tasks like simple instruction-following
databricks-gpt-5-3-codex	400,000	128,000	text, image	Most advanced agentic coding model, operating 25% faster
databricks-gpt-5-2-codex	400,000	128,000	text, image	Code-specialized LLM, excels at code generation (retiring July 16, 2026)

Details

Note: databricks-gpt-5-2-codex is listed as retiring on July 16, 2026 on the Databricks supported models page. It should be added with a deprecation_date field.

Suggested entries

"databricks-gpt-5-nano": {
  "format": "openai",
  "flavor": "chat",
  "multimodal": true,
  "displayName": "GPT-5 nano",
  "max_input_tokens": 400000,
  "max_output_tokens": 128000,
  "available_providers": ["databricks"]
},
"databricks-gpt-5-3-codex": {
  "format": "openai",
  "flavor": "chat",
  "multimodal": true,
  "displayName": "GPT-5.3 Codex",
  "max_input_tokens": 400000,
  "max_output_tokens": 128000,
  "available_providers": ["databricks"]
},
"databricks-gpt-5-2-codex": {
  "format": "openai",
  "flavor": "chat",
  "multimodal": true,
  "displayName": "GPT-5.2 Codex",
  "max_input_tokens": 400000,
  "max_output_tokens": 128000,
  "deprecation_date": "2026-07-16",
  "available_providers": ["databricks"]
}

Pricing omitted per Databricks provider constraints (no stable public per-model pricing source).

Verification checklist

• Cross-source: Model IDs confirmed on Foundation model overview and Supported models page
• Recent commits: Not covered by any recent commit to model_list.json
• ID format: Follows existing databricks-<model> naming convention
• Existing issues: Not covered by [BOT ISSUE] Databricks: add missing databricks-claude-opus-4-7, databricks-claude-sonnet-4-6, databricks-gpt-5-5 #559, [BOT ISSUE] Databricks: add missing databricks-gpt-5-5-pro, databricks-gemini-3-5-flash, databricks-gpt-5-4-mini #585, [BOT ISSUE] Databricks: add missing databricks-llama-4-maverick, databricks-gpt-oss-120b, databricks-gemma-3-12b #591, [BOT ISSUE] Databricks: add missing databricks-claude-sonnet-4-5, databricks-claude-opus-4-6, databricks-claude-sonnet-4 #592, [BOT ISSUE] Databricks: add missing databricks-gpt-5-4, databricks-gpt-5-4-nano, databricks-gpt-5-2 #638, [BOT ISSUE] Databricks: add missing databricks-gemini-3-1-pro, databricks-gemini-2-5-pro, databricks-gemini-2-5-flash #639, or [BOT ISSUE] Databricks: add missing databricks-claude-haiku-4-5, databricks-claude-opus-4-5, databricks-claude-opus-4-1 #640

Verification notes

Field	Source	Notes
Model IDs	Foundation model overview + Supported models page	Listed in both pay-per-token model tables
Context windows (400K)	Supported models page	All three: "400K total token context window with 128K maximum output tokens"
Max output (128K)	Supported models page	Same source as context window
format: "openai", flavor: "chat"	Catalog convention	All Databricks models use openai format
multimodal: true	Supported models page	Inputs: text, image
Retirement date for gpt-5-2-codex	Supported models page	"This model is retiring July 16, 2026"
input_cost_per_mil_tokens / output_cost_per_mil_tokens	NOT PUBLISHED	Databricks does not publish stable per-model pricing

Local files inspected

• packages/proxy/schema/model_list.json — grep for databricks-gpt-5-nano, databricks-gpt-5-3-codex, and databricks-gpt-5-2-codex returns no matches

Source URLs

• https://docs.databricks.com/aws/en/machine-learning/model-serving/foundation-model-overview
• https://docs.databricks.com/aws/en/machine-learning/foundation-model-apis/supported-models

{
  "kind": "missing_model",
  "provider": "databricks",
  "models": ["databricks-gpt-5-nano", "databricks-gpt-5-3-codex", "databricks-gpt-5-2-codex"],
  "status": "active",
  "model_specs": {
    "databricks-gpt-5-nano": {
      "format": "openai",
      "flavor": "chat",
      "multimodal": true,
      "displayName": "GPT-5 nano",
      "max_input_tokens": 400000,
      "max_output_tokens": 128000,
      "available_providers": ["databricks"]
    },
    "databricks-gpt-5-3-codex": {
      "format": "openai",
      "flavor": "chat",
      "multimodal": true,
      "displayName": "GPT-5.3 Codex",
      "max_input_tokens": 400000,
      "max_output_tokens": 128000,
      "available_providers": ["databricks"]
    },
    "databricks-gpt-5-2-codex": {
      "format": "openai",
      "flavor": "chat",
      "multimodal": true,
      "displayName": "GPT-5.2 Codex",
      "max_input_tokens": 400000,
      "max_output_tokens": 128000,
      "deprecation_date": "2026-07-16",
      "available_providers": ["databricks"]
    }
  },
  "source_urls": [
    "https://docs.databricks.com/aws/en/machine-learning/model-serving/foundation-model-overview",
    "https://docs.databricks.com/aws/en/machine-learning/foundation-model-apis/supported-models"
  ]
}

[github-actions]

Prepared catalog updates for databricks-gpt-5-nano, databricks-gpt-5-3-codex. Added models: databricks-gpt-5-nano, databricks-gpt-5-3-codex. Skipped deprecated or near-deprecation models: databricks-gpt-5-2-codex.

Verification sources: https://docs.databricks.com/aws/en/machine-learning/model-serving/foundation-model-overview, https://docs.databricks.com/aws/en/machine-learning/foundation-model-apis/supported-models
Verified metadata applied:

• databricks-gpt-5-nano: format=openai, flavor=chat, displayName=GPT-5 nano, max_input_tokens=400000, max_output_tokens=128000, pricing=not provided, providers=databricks, multimodal=true
• databricks-gpt-5-3-codex: format=openai, flavor=chat, displayName=GPT-5.3 Codex, max_input_tokens=400000, max_output_tokens=128000, pricing=not provided, providers=databricks, multimodal=true

Opened #652 to apply the catalog update.