← Founder Notes
Archive

Hydrafusion hit copilot research preview today and matched opus 5 at lower cost by routing between…

Yethikrishna ROriginal on Threads

hydrafusion hit copilot research preview today and matched opus 5 at lower cost by routing between models. the herd will call it orchestration. it is actually the end of the single-model moat.

which one of you is going to be the router, and which one of you is going to be the routed?

Context

GitHub's blog of 4 September 2026 says that in controlled offline evaluations, HydraFusion's selective coding workflows matched or exceeded the evaluated Opus 5 baseline while reducing estimated workflow cost. It is available as a research preview in GitHub Copilot on all plans through /experimental in the Copilot CLI, billed at each underlying model's standard rate.

How it compares

These are GitHub's own offline evaluations of selective workflows against a specific baseline, with an estimated cost. They are not an independent or general result. VentureBeat's headline says it cuts costs in every benchmark and matches quality in only one. The end of the single-model moat, and the router and routed question, are the author's opinion.

Related work

Watch next

  • Independent evaluations of HydraFusion.

Sources

  1. Project HydraFusion (GitHub blog)github.blog
  2. GitHub's HydraFusion cuts AI coding costs (VentureBeat)venturebeat.com

Provenance

The note above is reproduced unedited from the original post, first published on Threads on 4 September 2026 at 22:04 IST. Sources are the papers and datasets the note draws on.

View the original post
Embed this note
<iframe src="https://founder.myndlabs.tech/notes/embed/hydrafusion-hit-copilot-research-preview-today-and-matched-Dc3055IDQF-" width="480" height="420" style="border:0;max-width:100%" loading="lazy" title="Hydrafusion hit copilot research preview today and matched opus 5 at lower cost by routing between…"></iframe>

More notes