ABC-Bench: leaderboard
Agentic Bio-Capabilities Benchmark for Biosecurity. Measures LLM agents on end-to-end biosecurity-relevant tasks such as liquid-handling robot code, DNA fragment design, and synthesis-screening evasion.
Metric: Overall pass@1 (imported). Source: openreview.net. Source: benchmarklist.com. Saturation forecast: Around March 2027. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.5 | 63.2 | #142 |
| 2 | DeepSeek V3.2 | 50.1 | #205 |
| 3 | GPT-5 | 49.4 | #98 |
| 4 | Qwen 3 Coder 480B A35B | 43.1 | #279 |
| 5 | GLM-4.7 | 40.1 | #192 |
| 6 | Gemini 2.5 Pro | 25 | #152 |
| 7 | Qwen 3 32B | 8.9 | #443 |
| 8 | Qwen 3 8B | 8.3 | #674 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=abc-bench · How It Works · Data refreshed daily, snapshot 2026-10-20.