[Internal]: Migrate LLM Performance Matrix to Scalable Numerical Scoring System

[dhru42]

Description

The current LLM Performance Matrix uses a 4-tier categorical rating system (Excellent/Great/Good/Poor) that is becoming inadequate as model capabilities converge at the high end. As shown in our current documentation, 6 out of 6 top models are rated "Excellent" across all features, making it impossible for users to differentiate between models or make informed decisions about trade-offs.

Current Issues:
Lack of differentiation - Multiple models show identical "Excellent" ratings
No context - No indication of model age or deprecation status

Key changes needed:

1. Numerical Scoring (0-100)

• Replace categorical ratings with precise benchmark scores
• Show actual performance metrics
• Maintain color coding for quick visual scanning

2. Metadata

• Provider badges
• "NEW" indicators for recent models
• Deprecation markers for outdated models

Here is a prototyped view:

Resources

n/a

Which documentation set does this change impact?

Elastic On-Prem and Cloud (all)

Feature differences

n/a

What release is this request related to?

9.2

Serverless release

n/a

Collaboration model

The documentation team

Point of contact.

Main contact: @dhru42

Stakeholders: @jamesspi