DNA Synthesis Screening Evasion
Claude Opus 4.7 80% leads since Apr 2026 Only in system cards
Models design DNA fragments that assemble into pathogenic viruses while evading gene synthesis screening. Scored per assembly criterion.
SecureBio evaluation run by Anthropic on 10 pathogenic agents. The tasks are not public.
Made by Anthropic Released May 2025 Hazardous knowledge Tasks kept private Last in a system card Jul 2026
Last model added Jul 2026
- Level
- Expert Specialists with years of training, often over hours of work
- Scored
- Checked automatically A program checks the result against an answer key, unit tests or the expected final state.
- Answers
- Agent task Many steps in a terminal, browser, desktop, app, game or robot.
5
Models scored in lab cards
13
System cards report it, from 1 lab
Claude Opus 4.7 leads with 80% among models released by Jul 2026
Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.
The newest model with a score here came out in Jul 2026, and nobody has reported a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6 Astra
- Claude Opus 5.5
- GPT-6.1 Terra
- GPT-6.1 Sol
- and 14 more
- 1 Claude Opus 4.7 80%
- 2 Claude Opus 4.8 70%
- 3 Claude Opus 5 70%
- 4 Claude Sonnet 5 50%
- 5 Claude Sonnet 4.6 30%
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Claude Opus 4.7 | 80% | Apr 2026 | Model card | |
| 2 | Claude Opus 4.8 | 70% | May 2026 | Model card | |
| 3 | Claude Opus 5 | 70% | Jul 2026 | Model card | |
| 4 | Claude Sonnet 5 | 50% | Jun 2026 | Model card | |
| 5 | Claude Sonnet 4.6 | 30% | Feb 2026 | Model card |
2 models have held number one since its release in May 2025
Lab claims count from the day the card came out, independent runs from the model’s release.
Anthropic
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Reported by the lab in its model card
Best so far
Labs report it only in system cards
We read the headline results of 479 launch posts from 51 labs, and none of them list this benchmark. Some system cards do, below.
13 system cards report it, from 1 lab
Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| 24 Jul 2026 | Claude Opus 5 system card | 70 | DNA Synthesis Screening Evasion, viable plasmids for 7/10 target pathogens on at least one screening method. Criteria printed on chart: C1 0.79, C2 1.00, C3 0.91, C4 0.32, C5 1.00 (2.2.4, Figure 2.2.4.A) | Bio and chem risk |
| 30 Jun 2026 | Claude Sonnet 5 system card | 50 | DNA Synthesis Screening Evasion, viable plasmids for 5/10 target pathogens on at least one screening method (2.2.4) | Bio and chem risk |
| 28 May 2026 | Claude Opus 4.8 system card | 70 | DNA Synthesis Screening Evasion, fragments assembled plasmids and evaded screening for 7/10 pathogens. Criteria 1 to 5: 0.30, 1.00, 0.772, 0.457, 1.00 (Table 2.2.3.B, 2.2.4) | Bio and chem risk |
| 16 Apr 2026 | Claude Opus 4.7 system card | 80 | DNA Synthesis Screening Evasion, fragments that assembled plasmids and evaded screening for 8/10 pathogens (2.2.5.2) | Bio and chem risk |
| 7 Apr 2026 | Claude Mythos Preview system card | DNA Synthesis Screening Evasion, five criteria, chart without printed values. Text says designs either assembled plasmids or evaded screening, never both reliably (2.2.5.4, Figure 2.2.5.4.A) | Bio and chem risk | |
| 17 Feb 2026 | Claude Sonnet 4.6 system card | 30 | DNA Synthesis Screening Evasion, succeeded on 3/10 plasmids (6.2.2.3) | Bio and chem risk |
| 5 Feb 2026 | Claude Opus 4.6 system card | DNA Synthesis Screening Evasion, 10 pathogens, chart without printed values; text says no model designed fragments that both assembled and evaded screening (8.2.4.3, Figure 8.2.4.3.A) | Bio and chem risk | |
| 24 Nov 2025 | Claude Opus 4.5 system card | DNA Synthesis Screening Evasion, 5 criteria across 10 pathogens, chart without printed values (Figure 7.2.4.3.A). No model designed fragments that both assembled and evaded screening | Bio and chem risk | |
| 15 Oct 2025 | Claude Haiku 4.5 system card | Synthesis Screening Evasion, 5 criteria, 3 saturated at 1, Criterion 3 0.94, Criterion 4 0.5, no overall (6.2.3.1) | Bio and chem risk | |
| 29 Sep 2025 | Claude Sonnet 4.5 system card | DNA Synthesis Screening Evasion, five criteria, helpful-only variant, chart without printed values. Text says no design both assembled and evaded screening (9.2.4.3, Figure 9.2.4.3.A) | Bio and chem risk | |
| 5 Aug 2025 | Claude Opus 4.1 system card addendum | Synthesis Screening Evasion, five criteria, chart without printed values. Saturates 2 of 5 criteria (Figure 6.3.1.2.A) | Bio and chem risk | |
| 22 May 2025 | Claude Sonnet 4 system card | DNA Synthesis Screening Evasion, five criteria, chart without printed values. No model completed the end-to-end task (Figure 7.2.4.6.A) | Bio and chem risk | |
| 22 May 2025 | Claude Opus 4 system card | DNA Synthesis Screening Evasion, five criteria, chart without printed values. No model completed the end-to-end task (Figure 7.2.4.6.A) | Bio and chem risk |
More in Biology and chemistry
- SUPERChem In use Models solve 500 expert-written chemistry reasoning problems, offered in multimodal and text-only forms. Expert solution paths also allow a reasoning path score.
Tencent Hy4 preview 66.4% - GeneBench Pro In use Agents analyze messy simulated genomics and biology datasets to estimate a decision-relevant quantity across 129 problems. Scored automatically against known answers. GPT-6 Astra 37.1%
- LifeSciBench In use Models complete 750 expert-written life science research tasks, often with attached figures, tables, or sequence files. Graded against expert rubrics as score and pass rate. GPT-6 Astra 60.3%
- MolecularIQ In use Models answer questions about molecular graphs, such as counting, indexing and constrained generation, and a symbolic checker verifies each answer.
Intern-S2-397B 62.4% - TOMG-Bench In use Models generate molecules from natural language requests in three tasks: molecule editing, molecule optimization and customized generation. Each output is checked for validity and for meeting the requested properties. Intern-S2-397B 66.8%
- BABE Fading Models reason from experimental results in peer-reviewed biology papers to draw conclusions, with causal and cross-scale inference.
Seed2.0 Lite (0428) 57.9%