[BOT ISSUE] OpenAI: add missing gpt-rosalind-research

[github-actions]

Missing model

gpt-rosalind-research is not present in model_list.json.

Verified fields

Field	Value	Source
format	openai	Same as other OpenAI models in catalog
flavor	chat	Listed under text/chat on OpenAI models page
input_cost_per_mil_tokens	5	OpenAI pricing page
output_cost_per_mil_tokens	25	OpenAI pricing page
input_cache_read_cost_per_mil_tokens	0.50	OpenAI pricing page
reasoning	true	Built on GPT-5.5 (a reasoning model); described as a "frontier reasoning model" on openai.com/index/introducing-gpt-rosalind
available_providers	["openai"]	Listed on OpenAI models page

Not published (omitted rather than guessed)

• max_input_tokens: not documented on the models page or pricing page. GPT-5.5 (the base model) uses 1,050,000 but the Rosalind variant may differ.
• max_output_tokens: not documented publicly. GPT-5.5 uses 128,000 but cannot confirm for Rosalind.
• multimodal: not confirmed — the models page description does not specify image input support.

Cross-source verification

1. OpenAI models page (developers.openai.com/api/docs/models): Lists gpt-rosalind-research under "Specialized models → Life sciences" with description "Life sciences reasoning for approved organizations"
2. OpenAI pricing page (developers.openai.com/api/docs/pricing): Lists gpt-rosalind-research under "Specialized Text/Chat Models" at $5.00 input / $25.00 output / $0.50 cached input per million tokens
3. GPT-Rosalind-5.5 System Card (deploymentsafety.openai.com/gpt-rosalind-5-5): Confirms model is incrementally trained from GPT-5.5; billing begins October 5, 2026

Suggested catalog entry

"gpt-rosalind-research": {
  "format": "openai",
  "flavor": "chat",
  "input_cost_per_mil_tokens": 5,
  "output_cost_per_mil_tokens": 25,
  "input_cache_read_cost_per_mil_tokens": 0.5,
  "displayName": "GPT-Rosalind Research",
  "reasoning": true,
  "available_providers": ["openai"]
}

Notes

• GPT-Rosalind is OpenAI's frontier reasoning model for life sciences research, incrementally trained from GPT-5.5
• Access requires OpenAI's trusted access program (similar to gpt-5.6-cyber which is also restricted to authorized organizations and is present in the catalog)
• Billing at $5/$25 begins October 5, 2026; the model is currently in a free preview phase
• The pricing page notes "Cache-write pricing does not apply to this model"
• Partners include Amgen, Moderna, the Allen Institute, and Thermo Fisher Scientific

Local files inspected

• packages/proxy/schema/model_list.json — confirmed gpt-rosalind-research absent (grep returns no matches)
• packages/proxy/schema/index.ts — reviewed for naming conventions

{
  "kind": "missing_model",
  "provider": "openai",
  "models": ["gpt-rosalind-research"],
  "status": "active",
  "model_specs": {
    "gpt-rosalind-research": {
      "format": "openai",
      "flavor": "chat",
      "input_cost_per_mil_tokens": 5,
      "output_cost_per_mil_tokens": 25,
      "input_cache_read_cost_per_mil_tokens": 0.5,
      "displayName": "GPT-Rosalind Research",
      "reasoning": true,
      "available_providers": ["openai"]
    }
  },
  "source_urls": [
    "https://developers.openai.com/api/docs/models",
    "https://developers.openai.com/api/docs/pricing"
  ]
}

[github-actions]

Autofix could not safely resolve this ticket.

Autofix skipped because the issue does not yet contain enough verified metadata to safely apply catalog changes for gpt-rosalind-research. Refusing to add gpt-rosalind-research without verified token limits in the issue metadata

Verified information from this run:

• Provider: openai
• Parsed models: gpt-rosalind-research
• Official source URLs already on the ticket: https://developers.openai.com/api/docs/models, https://developers.openai.com/api/docs/pricing, https://openai.com/index/introducing-gpt-rosalind/, https://deploymentsafety.openai.com/gpt-rosalind-5-5

To make autofix work on a future run, update the machine-readable metadata block for this missing model issue and include the verified fields needed for the catalog change, such as:

• model_spec for single-model issues or model_specs for multi-model issues
• displayName
• parent for non-base dated snapshots or location-scoped variants when official sources verify the relationship
• pricing fields
• max_input_tokens
• max_output_tokens
• available_providers
• deprecated / deprecation_date
• locations when the provider/model requires explicit location metadata