ResearchBenchmarksMeeting claim evaluation on city_council & private_data overall (test)Follow75.4Accuracygpt-5.473.8474.24574.6575.055Apr 23, 2026Evaluation ResultsMethodMethodLinksAccuracyCompletenessCoveragegpt-5.4Meetings=56, Evaluator...Meetings=56, Evaluator runs=1122026.0475.484.486.8gpt-4.1Meetings=56, Evaluator...Meetings=56, Evaluator runs=1122026.0473.978.979.4