The AI arena is free today

Open Superagent

Labwork Bench

Paper

Progress Over Time

Interactive timeline showing model performance evolution on Labwork Bench

State-of-the-art frontier
Open
Proprietary

Labwork Bench Leaderboard

1 models
ContextCostLicense
1——$5.00 / $25.00
Notice missing or incorrect data?
About this benchmark

What is Labwork Bench?

Labwork Bench is an OpenAI evaluation of whether models can connect perturbations to outcomes in real wet-lab protocols for troubleshooting and optimization tasks.

Labwork Bench is a text benchmark evaluating models on reasoning, science, healthcare, agents, and biology tasks. LLM Stats tracks 1 models on this benchmark, scored on a 0–1 scale. The current average is 0.6, with the leader at 0.6.

Compare leaders on the best AI for reasoning, best AI for science, best AI for healthcare, best AI for agents and best AI for biology leaderboards.

Current leaders

GPT-Rosalind from OpenAI currently leads the Labwork Bench leaderboard with a score of 0.632 across 1 evaluated AI models.

1GPT-RosalindOpenAI63.2%

FAQ

Common questions about the Labwork Bench benchmark and leaderboard.

What is the Labwork Bench benchmark?

Labwork Bench is an OpenAI evaluation of whether models can connect perturbations to outcomes in real wet-lab protocols for troubleshooting and optimization tasks.

What is the Labwork Bench leaderboard?

The Labwork Bench leaderboard ranks 1 AI models based on their performance on this benchmark. Currently, GPT-Rosalind by OpenAI leads with a score of 0.632. The average score across all models is 0.632.

What is the highest Labwork Bench score?

The highest Labwork Bench score is 0.632, achieved by GPT-Rosalind from OpenAI.

How many models are evaluated on Labwork Bench?

1 models have been evaluated on the Labwork Bench benchmark, with 0 verified results and 1 self-reported results.

Where can I find the Labwork Bench paper?

The Labwork Bench paper is available at https://deploymentsafety.openai.com/gpt-rosalind-5-5. The paper details the methodology, dataset construction, and evaluation criteria.

What categories does Labwork Bench cover?

Labwork Bench is categorized under reasoning, science, healthcare, agents, and biology. The benchmark evaluates text models.

Which model offers the best value on Labwork Bench?

Among models scoring within 10% of the leader, GPT-Rosalind from OpenAI is the cheapest, at $5.00 per million input tokens with a score of 0.632.

How recent are the Labwork Bench leaderboard results?

The Labwork Bench leaderboard was last updated in November 2026 and currently includes 1 evaluated models.