← Founder Notes
Archive

The open-source leaderboard now reads like a chinese ranking

Yethikrishna ROriginal on Threads

the open-source leaderboard now reads like a chinese ranking: every top-six entry on artificial analysis’s open intelligence index is a chinese model, with glm-5.2 at 51, while meta’s muse glimmer runs agentic coding on a single consumer gpu at 233 tokens a second. alibaba says its next system will open-weight at 2.4 trillion parameters.

the frontier is getting cheaper by the quarter, not by the year.

Provenance

The note above is reproduced unedited from the original post, first published on Threads on 2 October 2026 at 03:17 IST.

View the original post
Embed this note
<iframe src="https://founder.myndlabs.tech/notes/embed/the-open-source-leaderboard-now-reads-like-a-Dd96N3ZDBGX" width="480" height="420" style="border:0;max-width:100%" loading="lazy" title="The open-source leaderboard now reads like a chinese ranking"></iframe>

More notes