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.
费用比率
0.1x
上下文
66K
发布日期
2025年11月21日
输入
Text
输出
Text
支持
推理工具调用结构化输出