Anthropic: Claude designed protein binders for 14 of 15 targets, beating industry success rates
FinanceYF5 · x · 2026-08-19
Anthropic's blog shares two life-science results:
- Protein binder design: Claude (Mythos Preview and Opus 4.8) designed de novo protein binders against 15 targets, succeeding on 14. Individual design success rates of 22%–35% compare to a 10%–15% industry norm; some strongest designs bound several times more tightly than the best published results—a task that historically took specialists weeks to months per target.
- Chemical analysis: Given a contract lab's raw NMR and LC-MS files plus a two-sentence prompt, Claude Opus 5 returned finished analyses in 23 and 19 minutes, matching the lab's own results on hydrogen counts and purity (96.4% vs 96.33%).
Anthropic adds that a scientist access program remains a top priority with details coming soon, and that Opus 5 is currently the best Claude model for life-science research.
Related event: Anthropic: Claude Autonomously Designs Proteins, Hitting 14 of 15 Targets(28 posts)→
More from Models
- ChatGPT defaults to French when reasoning level is set low — StephanSturges · 2026-08-19
- Qwen 3.8 27B is missing a 'high' reasoning effort between medium and xhigh — alanoo · 2026-08-19
- 'Scaling laws are not laws of nature': better data and architectures can still bend the curves — bookwormengr · 2026-08-19
- Facebook Releases MoE-ViE: Efficient Mixture-of-Experts Vision Encoder — facebook · 2026-08-19
- Qwen3.8-27B users report grammar errors in non-English outputs — LegacyRemaster · 2026-08-19
- DeepSeek Web Chat Wins Industrial Track in TAAC × KDD Cup — jiqizhixin · 2026-08-19