Benchmarks.bio - TxBench-AD: leaderboard
Antibody discovery run end to end: 100 agentic evaluations spanning ten stages from target hypothesis and antigen design through binder generation, epitope and escape analysis, to preclinical candidate de-risking. Each starts from real experimental data captured just before an analysis step and is graded deterministically on whether the agent reaches the actual result.
Metric: Pass Rate (%). Source: benchmarks.bio. Saturation forecast: Around October 2027. 14 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6 Astra | 56.9 | #6 |
| 2 | Claude Opus 5 | 52.96 | #12 |
| 3 | Grok 4.6 | 51.85 | #26 |
| 4 | Gemini 3.7 Flash | 50 | #23 |
| 5 | Claude Opus 4.8 | 45.64 | #37 |
| 6 | Gemini 3.5 Flash | 42.62 | #60 |
| 7 | Claude Opus 4.7 | 39.03 | #49 |
| 8 | DeepSeek V4.1 Flash | 39 | #36 |
| 9 | Grok 4.5 | 38.85 | #48 |
| 10 | Claude Sonnet 5 | 37.26 | #57 |
| 11 | GPT-5.6 Sol | 33.78 | #19 |
| 12 | GPT-5.5 | 33.56 | #31 |
| 13 | GPT-5.6 Terra | 30.98 | #41 |
| 14 | GPT-5.6 Luna | 19.39 | #68 |
No result here: #3 Claude Opus 5.5, #7 GPT-6.1 Sol, #8 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=benchmarks-bio-txbench-ad · How It Works · Data refreshed daily, snapshot 2026-10-20.