Skip to content

More detailed stats, removed incompatible models#208

Merged
georgeh0 merged 6 commits into
cocoindex-io:mainfrom
jimmckeeth:main
Jul 2, 2026
Merged

More detailed stats, removed incompatible models#208
georgeh0 merged 6 commits into
cocoindex-io:mainfrom
jimmckeeth:main

Conversation

@jimmckeeth

Copy link
Copy Markdown
Contributor

Me again. I've been investigating the different models, did some bench marking and removed a reference to an incompatible model. There are some really good options listed in there, like lightonai/LateOn-Code-edge - it is much faster than the current default, even smaller, and with a better MTEB ranking.

Tier Model Params Dims Code Score CPU Speed Best For
Default Snowflake/arctic-embed-xs 22M 384 0.67 Very Fast Smallest, most compatible default
Micro lightonai/LateOn-Code-edge 17M 256 0.82 Crazy Fast (231 cps) Efficiency King. Incredible code performance for its size.
Small lightonai/LateOn-Code 149M 768 0.85 Fast (7 cps) Great balance of speed and accuracy on modern laptops.
Medium Shuu12121/CodeSearch-ModernBERT-Crow-Plus 152M 0.89 Fast High accuracy with encoder speed; best for CPU-only indexing at this tier.

@jimmckeeth jimmckeeth marked this pull request as ready for review July 1, 2026 06:19
@badmonster0 badmonster0 requested a review from georgeh0 July 1, 2026 19:09
@badmonster0

Copy link
Copy Markdown
Member

@jimmckeeth this is great, thanks a lot! @georgeh0 can help with review!

@georgeh0 georgeh0 left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks a lot!

@georgeh0 georgeh0 merged commit e709902 into cocoindex-io:main Jul 2, 2026
4 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants