Reports

Monthly LLM benchmark reports

What each month's runs showed, written up by Suhas Bhairav: cost per correct answer, what reasoning tokens cost and buy, the fastest models and how far open-weights models trail. Each report is fixed to the last run of its month.