Add frontier AI models value set with hierarchical taxonomy

[cmungall]

Summary

Adds a comprehensive hierarchical value set for frontier large language and foundation models, organized by provider, model family, and specific release versions. This new schema module enables standardized representation of current-generation AI models across major developers.

Changes

• New schema module: src/valuesets/schema/computing/frontier_ai_models.yaml

  • Defines FrontierModelEnum with three-tier hierarchy: developer → family → specific model
  • Covers major providers: Anthropic (Claude), OpenAI (GPT), Google DeepMind (Gemini), xAI (Grok), and open-weight models (Llama, DeepSeek, Mistral, Kimi)
  • Includes 50+ model variants with canonical API identifiers
  • Annotates models with metadata: context window, max output tokens, pricing, training cutoff, release dates, and links to model cards
  • Uses node_type annotations to distinguish developer/family/model tiers

• Updated main schema: src/valuesets/schema/valuesets.yaml

  • Added import for the new computing/frontier_ai_models module

Implementation Details

• Hierarchical structure uses is_a relationships to organize models under families and developers
• Leaf nodes use verbatim API model identifiers (e.g., claude-opus-4-8, gpt-5.5, gemini-3.5-flash)
• Intermediate categorical nodes use uppercase naming (e.g., ANTHROPIC, CLAUDE_OPUS)
• Metadata captured as annotations and see_also links where available
• Status marked as DRAFT to reflect the rapidly evolving frontier model landscape (snapshot as of mid-2026)
• Includes contributor attribution via ORCID

https://claude.ai/code/session_01A9nEWUPAsNgqqAjLoaG1ny

[claude]

Claude finished @cmungall's task —— View job

---

PR Review: Frontier AI Models Value Set

Reviewing this pull request for code quality, potential issues, and best practices compliance.

Todo List:

• Read project guidelines (CLAUDE.md)
• Examine the new frontier AI models schema file
• Review the main schema import addition
• Validate LinkML schema structure and conventions
• Check for potential bugs or issues
• Assess security and privacy considerations
• Evaluate test coverage needs
• Provide comprehensive review feedback

Review Summary

This is an excellent addition to the valuesets project! The frontier AI models schema is well-structured, follows LinkML best practices, and provides comprehensive coverage of current frontier models. Here's my detailed analysis:

✅ Strengths

1. Excellent Schema Structure

• Perfect adherence to the project's LinkML patterns and conventions
• Proper hierarchical organization using is_a relationships (developer → family → model)
• Consistent naming: uppercase for categories, lowercase API identifiers for leaf nodes
• Well-documented with clear descriptions and comprehensive metadata

2. Strong Ontological Design

• Smart use of node_type annotations to distinguish hierarchy levels (developer, family, model)
• Comprehensive metadata capture: context windows, pricing, training cutoffs, release dates
• Appropriate use of see_also links to official documentation
• Good use of aliases for common model nicknames

3. Project Standards Compliance

• Follows CLAUDE.md guidelines: CamelCase enum names, proper CURIE usage
• Correct import structure and prefix declarations
• Proper contributor attribution with ORCID
• Status marked as DRAFT (appropriate for rapidly evolving landscape)

4. Comprehensive Coverage

• Covers all major frontier providers: Anthropic, OpenAI, Google DeepMind, xAI
• Includes open-weight models: Meta Llama, DeepSeek, Mistral, Moonshot
• Future-compatible structure ready for new models and providers

🔍 Observations & Suggestions

1. Metadata Consistency src/valuesets/schema/computing/frontier_ai_models.yaml:77-85
Some models have complete pricing metadata while others don't. Consider:

# Example: claude-opus-4-7 is missing pricing info
claude-opus-4-7:
  # Consider adding pricing metadata if available
  annotations:
    input_usd_per_mtok: 5.00  # if known
    output_usd_per_mtok: 25.00  # if known

2. API Identifier Validation
The schema uses canonical API identifiers - excellent! But consider adding validation that these match actual provider APIs as they evolve.

3. Future Versioning Strategy
The use of snapshot dates (mid-2026) in descriptions is good. Consider adding a data_version or snapshot_date annotation at the enum level for tracking.

🛡️ Security & Privacy

• ✅ No security concerns - contains only public model metadata
• ✅ No PII or sensitive information exposed
• ✅ All references are to publicly documented APIs

📋 Minor Recommendations

1. Model ID Consistency: Consider adding full_model_id annotations consistently where providers use them (some models have this, others don't)

2. License Information: Add license annotations for open-weight models where missing (DeepSeek, Moonshot don't have license info)

3. Release Date Standardization: Consider ISO 8601 format for release dates (2026-05-19 vs 2025-12)

🧪 Testing Considerations

• The schema structure looks solid for LinkML validation
• Consider adding example usage in tests/data/valid/
• Validation commands should pass cleanly with the current structure

📊 Performance & Maintainability

• Excellent organization will make updates easy as new models are released
• Clear hierarchy makes programmatic traversal straightforward
• Good foundation for automated updates from provider APIs

Overall Assessment: EXCELLENT ✨

This PR represents high-quality schema design that:

• Follows all project conventions perfectly
• Provides comprehensive, well-organized coverage
• Uses LinkML features effectively
• Sets up an excellent foundation for future growth

The hierarchical structure, metadata richness, and attention to current provider ecosystems make this a valuable addition to the valuesets collection.

Recommendation: Approve and merge 🚀

---

[claude]

Claude finished @cmungall's task —— View job

---

****

PR Review: Frontier AI Models Value Set

Reviewing this pull request for code quality, potential issues, and best practices compliance.

Todo List:

• Read project guidelines (CLAUDE.md)
• Examine the new frontier AI models schema file
• Review the main schema import addition
• Validate LinkML schema structure and conventions
• Check for potential bugs or issues
• Assess security and privacy considerations
• Evaluate test coverage needs
• Provide comprehensive review feedback

Review Summary

This is an excellent addition to the valuesets project! The frontier AI models schema is well-structured, follows LinkML best practices, and provides comprehensive coverage of current frontier models. Here's my detailed analysis:

✅ Strengths

1. Excellent Schema Structure

• Perfect adherence to the project's LinkML patterns and conventions
• Proper hierarchical organization using is_a relationships (developer → family → model)
• Consistent naming: uppercase for categories, lowercase API identifiers for leaf nodes
• Well-documented with clear descriptions and comprehensive metadata

2. Strong Ontological Design

• Smart use of node_type annotations to distinguish hierarchy levels (developer, family, model)
• Comprehensive metadata capture: context windows, pricing, training cutoffs, release dates
• Appropriate use of see_also links to official documentation
• Good use of aliases for common model nicknames

3. Project Standards Compliance

• Follows CLAUDE.md guidelines: CamelCase enum names, proper CURIE usage
• Correct import structure and prefix declarations
• Proper contributor attribution with ORCID
• Status marked as DRAFT (appropriate for rapidly evolving landscape)

4. Comprehensive Coverage

• Covers all major frontier providers: Anthropic, OpenAI, Google DeepMind, xAI
• Includes open-weight models: Meta Llama, DeepSeek, Mistral, Moonshot
• Future-compatible structure ready for new models and providers

🔍 Observations & Suggestions

1. Metadata Consistency src/valuesets/schema/computing/frontier_ai_models.yaml:77-85
Some models have complete pricing metadata while others don't. Consider:

# Example: claude-opus-4-7 is missing pricing info
claude-opus-4-7:
  # Consider adding pricing metadata if available
  annotations:
    input_usd_per_mtok: 5.00  # if known
    output_usd_per_mtok: 25.00  # if known

2. API Identifier Validation
The schema uses canonical API identifiers - excellent! But consider adding validation that these match actual provider APIs as they evolve.

3. Future Versioning Strategy
The use of snapshot dates (mid-2026) in descriptions is good. Consider adding a data_version or snapshot_date annotation at the enum level for tracking.

🛡️ Security & Privacy

• ✅ No security concerns - contains only public model metadata
• ✅ No PII or sensitive information exposed
• ✅ All references are to publicly documented APIs

📋 Minor Recommendations

1. Model ID Consistency: Consider adding full_model_id annotations consistently where providers use them (some models have this, others don't)

2. License Information: Add license annotations for open-weight models where missing (DeepSeek, Moonshot don't have license info)

3. Release Date Standardization: Consider ISO 8601 format for release dates (2026-05-19 vs 2025-12)

🧪 Testing Considerations

• The schema structure looks solid for LinkML validation
• Consider adding example usage in tests/data/valid/
• Validation commands should pass cleanly with the current structure

📊 Performance & Maintainability

• Excellent organization will make updates easy as new models are released
• Clear hierarchy makes programmatic traversal straightforward
• Good foundation for automated updates from provider APIs

Overall Assessment: EXCELLENT ✨

This PR represents high-quality schema design that:

• Follows all project conventions perfectly
• Provides comprehensive, well-organized coverage
• Uses LinkML features effectively
• Sets up an excellent foundation for future growth

The hierarchical structure, metadata richness, and attention to current provider ecosystems make this a valuable addition to the valuesets collection.

Recommendation: Approve and merge 🚀

---