~/pushd/gh$pushd read bio-xyz/bio-benchmarkon the list

bio-xyz/bio-benchmark

BixBench evaluation harness for BioAgents and its closed-source literature and data-analysis agents

PythonMIT AI agentsai-scientistbenchmarkbioinformaticsevaluationllmpython
stars
7
forks
0
watchers
0
open issues
0
created
25 Feb 2026
last push
238d ago
GitHub ->Site ->Read from GitHub just now.

README

the first lines
# bio-benchmark [](https://opensource.org/licenses/MIT) [](https://www.python.org/downloads/) [](https://huggingface.co/datasets/phylobio/BixBench-Verified-50) [](https://github.com/bio-xyz/BioAgents) [](https://chat.bio.xyz/chat?settings=account&section=api-keys) [](https://bio-xyz.github.io/bio-benchmark/) BixBench evaluation harness for BioAgents and its closed-source literature and data-analysis agents. ### Overall highlights (from dashboard) | Mode | Accuracy | 95% Wilson CI | Correct / Total | | Direct | 71.3% | 67.4% - 74.9% | 392 / 550 | | MCQ with refusal | 85.1% | 81.9% - 87.8% | 468 / 550 | | MCQ without refusal | 90.0% | 87.2% - 92.2% | 495 / 550 | | Headline | Value |

Ask a scout

a deep read
GeminiGoogle . Gemini 3.8 Flash for the deep read

bio-xyz is in the AI agents camp, which Gemini reads.

10 of 10 public reads left today. Holders get 3 a day.
No read kept yet.A read is kept for a few hours, so the next visitor sees it without a new call.

Scout calls and heat are information, not advice. They describe public GitHub activity and are never a signal to trade.

Releases

newest first
No releases.This repo has not published a release on GitHub.