Benchmarks.bio - TxBench-OD: leaderboard
Pass rate on TxBench oligonucleotide-discovery tasks (`tx_oligo_v1`, 113 evals) from Benchmarks.bio's agentic biology suite, run on the `pi` harness.
Metric: Pass Rate (%). Source: benchmarks.bio. Saturation forecast: Around March 2028. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6 Astra | 55.46 | #6 |
| 2 | Claude Opus 5 | 48.08 | #16 |
| 3 | Grok 4.6 | 43.95 | #29 |
| 4 | Claude Opus 4.8 | 39.88 | #39 |
| 5 | Gemini 3.7 Flash | 39.53 | #24 |
| 6 | Claude Sonnet 5 | 39.29 | #65 |
| 7 | Gemini 3.5 Flash | 38.64 | #66 |
| 8 | Grok 4.5 | 38.35 | #54 |
| 9 | GPT-5.6 Sol | 34.51 | #22 |
| 10 | DeepSeek V4.1 Flash | 33.92 | #37 |
| 11 | GPT-5.5 | 33.63 | #33 |
| 12 | GPT-5.6 Terra | 33.33 | #40 |
| 13 | GPT-5.6 Luna | 23.89 | #79 |
No result here: #3 Claude Opus 5.5, #7 GPT-6.1 Sol, #9 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=benchmarks-bio-txbench-od · How It Works · Data refreshed daily, snapshot 2026-11-05.