Why We Trimmed the Model Zoo
By Team Valor AI · December 12, 2024
Valor used to carry a sprawling list of language models. It slowed down decisions and created uneven experiences. So we trimmed the zoo and doubled down on the options that consistently serve veterans best.
Highlights
- Removed redundant models that overlapped in capability but added latency or cost.
- Tiered lineup now aligns to core use cases: guided forms, research, and outreach.
- Pricing clarity makes it obvious how model choice affects workspace spend.
- Governance review ensures every remaining model meets compliance and safety bars.
Why it matters
Fewer, better choices mean faster onboarding and less second-guessing. Veterans get reliable guidance without the team juggling experimental settings.
How to explore it
- Open the workspace settings to review the current model catalog.
- Compare latency and cost benchmarks provided for each remaining model.
- Read the model policy to understand how we evaluate new additions.
Screenshots to capture
- Model selection screen with the simplified lineup.
- Tooltip or modal explaining pricing and recommended use cases.
- Governance checklist or policy document snippet.