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.
비용 비율
0.1x
컨텍스트
66K
출시일
2025년 11월 21일
입력
Text
출력
Text
지원
추론도구 호출구조화된 출력