Loading…
RadixArk’s open-weight agent-focused fine-tune reports a substantial Terminal-Bench gain, but independent reproduction and the generality of its efficiency claims are not established here.
Limits of the evidence: The performance and training-efficiency figures are reported in the release post; independent confirmation is not provided in the supplied text.
What to watch
Assessment revised Nov 18, 2026
DaisyRose-27B: Pushing the Frontier of Post-Training Efficiency
Hugging Face BlogNov 18, 2026First report
First assessment, from Hugging Face Blog
No maturity ladder applies to this desk.
Sens.ai aggregates and assesses published reporting. The assessment above is machine generated; the original sources are authoritative.