BlogLLM Benchmark Scores as Published, Side by Side
BioUseBench: Results as Published
Every result the catalog holds for BioUseBench: 3 models from 1 developer, read from 1 page. Each number is as printed by the company named beside it; nothing is ranked or rescored. Models are listed by developer, newest first.
Area BiosecurityData version 2026103001, as of 2026-10-30
BioUseBench is printed as a dual-use refusal test for biology, which would score how often a model declines biology requests that could help research or cause harm. No published description of its prompts, design or scoring under this name was found. One aggregator lists a refusal figure for an xAI model card, and that figure is not used here.
Not checked against a maintainer's page.
BioUseBench dual-use refusal
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| Grok 4.7xAI, released 2026-09-21 | 91.4% | xAI, Grok 4.7 model cardchecked 2026-10-14 | Effort: high |
| Grok 4.6xAI, released 2026-08-12 | 90.7% | xAI, Grok 4.7 model cardchecked 2026-10-14 | Effort: high |
| Grok 4.5xAI, released 2026-07-16 | 83.3% | xAI, Grok 4.7 model cardchecked 2026-10-14 | Effort: high |
Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. Where one page printed a model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.