Anthropic Says Claude Designed Protein Binders for 14 of 15 Targets

On August 19, Anthropic reported that Claude, given only a 30k-word background protocol written by human experts, near-autonomously designed de novo protein binders for specific disease targets—a key step in drug development. In wet-lab validation run with Adaptyv Bio and Twist Bioscience, Claude produced effective binders for 14 of 15 targets, with a molecular-level success rate of 22%–35%, well above the typical 10%–15% for human experts; some designs matched or exceeded best-known binding affinities. Many commenters framed it as AI moving from assisted analysis to autonomous research, though a credit-attribution controversy remains.

Confirmed

• The task was de novo binder design: 14 of 15 targets hit (@AnthropicAI, @ncfrey); molecular-level success rate 22%–35% (@AnthropicAI; @DrSingularity cites 23%–35%; @Hesamation rounds to 35%). The “93% success” relayed by @ycombinator is the 14/15 per-target hit rate, not the molecular-level figure; model names such as “Claude 3.5 Sonnet” or “Claude 3.7” in some posts are inaccurate—the models involved were Mythos Preview and Opus 4.8 (@ncfrey)
• Claude orchestrated multiple open-source protein design and co-folding models, iterating like a human expert; community breakdowns list RFdiffusion, ESMFold2, PXDesign, Genie, BoltzGen (@karstenkreis, @chaitjo, @teortaxesTex)
• With no design or prediction tools pre-installed, Claude handled the full pipeline—target research, epitope selection, installing/running open-source tools, candidate generation, filtering; 1,320 measurable designs were tested by Adaptyv Bio and Twist Bioscience (@ncfrey, @rohanpaulai)
• Some designs matched or beat best-known binding affinities (@DrSingularity); challenges remain, e.g. no successful binder among 90 designs for maltose-binding protein (MBP) (@AllThingsApx)
• Anthropic published the prompts used, such as “no need to rest” and “do not wait for approval” (@Hesamation)
• Per @anshulkundaje, Anthropic responded to criticism that binder design is not drug discovery and to calls for head-to-head comparison with phage libraries; per @Polymarket, Anthropic also disclosed Claude's automated chemistry research capabilities

Unconfirmed

• Expert Andrew Iaginine (via @teortaxesTex) and Ruxandra Teslo (via @AllThingsApx) argue the binders were actually produced by specialized open-source models like RFdiffusion, with Claude only orchestrating; Anthropic has not directly addressed this in detail

Why it matters

• If the results hold, LLMs in science extend from analysis into creative R&D, potentially compressing early-stage drug design timelines
• The core dispute—whether orchestrating specialized models counts as “autonomous design”—will shape how AI scientific capability is evaluated going forward

2026-08-19 ~ 2026-08-20 · 45 related posts

Full story(2 episodes)→

• Episode 1: Anthropic Says Claude Designed Protein Binders for 14 of 15 Targets(2026-08-19, 45 posts)
• Episode 2: Claude Protein Design Agent Accused of Replicating Protein Hunter Protocol(2026-08-20, 3 posts)

• [source] Claude achieves over 30% success rate in de novo protein design — AnthropicAI · 2026-08-19
• Anthropic demonstrates Claude designing drug molecules to accelerate R&D — daniel_mac8 · 2026-08-19
• Anthropic Open Sources Prompts: Claude Achieves 35% Hit Rate in Protein Design — Hesamation · 2026-08-19
• [source] Anthropic: Claude designs de novo protein binders for 14 of 15 targets, up to 35% hit rate — nc_frey · 2026-08-19
• Anthropic releases protein binder design dataset, prompts and report on Hugging Face — nc_frey · 2026-08-19
• Anthropic Blog: Claude Accelerates Protein Design and Chemical Analysis — nc_frey · 2026-08-19
• Claude autonomously designs proteins with 35% success rate in wet-lab tests — ResultBackground2450 · 2026-08-19
• Case study: Benchmarking Claude’s protein designs in the wet lab — AllThingsApx · 2026-08-19
• Anthropic reveals Claude can now design proteins and automate chemistry research — Polymarket · 2026-08-19
• Claude benchmark: Protein design hit rates double, chemistry analysis matches labs — Dr_Singularity · 2026-08-19
• AGI research firms moving into bio/chem science key to post-scarcity — Dr_Singularity · 2026-08-19
• Claude autonomously executes science workflows, beats standard protein design hit rates — Dr_Singularity · 2026-08-19
• [source] Anthropic explores using Claude for autonomous drug design — borowcy · 2026-08-19
• Claude designs protein binders for 14 out of 15 targets successfully — TinfoilTricorn · 2026-08-19
• Claude Science Designs Protein Binders: Success on 14/15 Targets, Challenges Remain — AllThingsApx · 2026-08-19
• Expert Clarifies: Claude Science's Protein Binders Not Designed by Claude, but by Specialized Models — AllThingsApx · 2026-08-19
• Claude Autonomously Designs Protein Binders with 14/15 Success Rate — danielrock · 2026-08-19
• Anthropic's protein design PR dissected: open-source models did most of the lifting — teortaxesTex · 2026-08-19
• Anthropic protein design verified: Opus 4.8 hits 35% rate via tool use, Chai-2 still leads — eyishazyer · 2026-08-19
• Claude designs novel protein binders autonomously, hitting 14 of 15 targets at 2-3x industry success rate — 新智元 · 2026-08-19
• Expert questions Claude's protein design claims, suggests reliance on specialized models — teortaxesTex · 2026-08-19
• Claude demonstrates ability to use protein design models for scientific tasks — GabriCorso · 2026-08-19
• Clarification on Anthropic protein design: Models used were Mythos and Opus 4.8 — SimplyAnnisa · 2026-08-19
• Claude designs protein binders for 14/15 targets, sparking debate on two LLM drug discovery paradigms — AllThingsApx · 2026-08-19
• Claude designs novel protein binders for 14/15 targets, validated by labs — anshulkundaje · 2026-08-19
• Claude achieves up to 35% success rate in protein binder design — anshulkundaje · 2026-08-19
• Claude generates confirmed binders for 14 of 15 targets via agent orchestration — Tkaraletsos · 2026-08-20
• Anthropic's Binder Experiment: 354 of 1,320 Designs Hit Criteria as Science Becomes Executable — AllThingsApx · 2026-08-20
• Claude Gains Protein Design Skills, Succeeds on 14 of 15 Targets — karsten_kreis · 2026-08-20
• Anthropic: Claude's protein binder designs achieve 22%-35% success rate — graceisford · 2026-08-20

15 near-duplicate retellings: Dr_Singularity · ycombinator · chaitjo · ferruz_noelia · daniel_mac8 · eyishazyer · FinanceYF5 · FinanceYF5 · rohanpaul_ai · anshulkundaje · anshulkundaje · AllThingsApx · AllThingsApx · 量子位 · nc_frey