Hello! Just wanted to share some major improvements we’ve made to Elicit this week:

🏆 Claude Opus 4.5 is live in Elicit

Claude Opus 4.5 was released three days ago, achieving state-of-the-art performance in tool use and novel problem solving, among other skills.

Our evaluations team benchmarked Claude Opus 4.5 against other frontier models like Claude Sonnet 4.5, Google's Gemini 3 Pro, and OpenAI's GPT 5 on extracting answers from papers and writing reports. We found that Opus is significantly better at data extraction and writing concise reports with fewer hallucinations.

Elicit for Pro, Teams, and Enterprise now uses Opus. This model is new and has limited capacity, so there will be a delay before it is rolled out to everyone else. 

🚀 Elicit Reports are 2x faster 

The average Elicit Report now generates in 4 minutes. Even comprehensive reports that analyze 1,000 papers should run within 15 minutes.

📝 Elicit Reports now use more prose instead of bullets and study counts

Previously, Elicit Reports focused on explaining how many studies followed a certain methodology, or arrived at a certain conclusion. This style was conducive to systematic reviews but not to other types of research. We’ve updated Elicit Reports to write in a more prose-based, narrative style, like this: