RFD4-Proteina benchmarks show step-change gains in protein binder design pass rates
AllThingsApx · x · 2026-10-21
Reviewer RafiBrent says that despite trying to remain skeptical, RFD4-Proteina represents a "step-change improvement" for in silico biomolecular design. Benchmark charts show: in protein–protein binder design, RFD4-Proteina (high thinking) achieves 10% clustered pass rate, far above BindCraft, Protein-Hunter, and RFdiffusion; 22/24 tasks have at least one passing design, and 66.5 mean unique passers per GPU-hour vs 9.6 for RFID3 — at an order of magnitude higher throughput.
Related event: RFD4-Proteina launches, beating baselines across protein design benchmarks(6 posts)→
More from Research
- New paper builds "bottleneck pressure" metric from 288M job postings to value work AI can't do — soumitrashukla9 · 2026-10-22
- Deep dive: running RL on heterogeneous hardware — Blackwells for training, TPUs for inference — A_K_Nain · 2026-10-22
- Long Now Labs picks first awardees, including a thermodynamic clock experiment for LLMs — begusgasper · 2026-10-22
- Princeton opens AI Postdoc fellowship in Natural and Artificial Minds, deadline Feb 28 — LakeBrenden · 2026-10-22
- ESMFold2 API adds max accuracy mode after single-sequence precision complaints — proteinrosh · 2026-10-22
- OpenTac team to demo tactile sensors at CoRL with small-scale sensor giveaway — anand_bhattad · 2026-10-22