OLMo 3 32B Think is Allen Institute for AI's fully open 32B reasoning model, released with open weights, training data, and code so its chain-of-thought can be inspected and traced back to training decisions. It is tuned for math and coding reasoning via long chain-of-thought, though it lacks tool use or vision support, making it best suited for transparent, auditable reasoning tasks rather than agentic workflows.
Тариф стоимости
0.1x
Контекст
66K
Выпуск
21 нояб. 2025 г.
Ввод
Text
Вывод
Text
Поддержка
РассуждениеВызов инструментовСтруктурированный вывод