Ling 3.0 Tiny
InclusionAI🇨🇳
inclusionai/ling-3.0-tinyLing 3.0 Tiny is InclusionAI's (Ant Group) smallest Ling 3.0 model, an MIT-licensed hybrid-linear mixture-of-experts model (7.9B total, 1.3B active parameters, 262K-token context) that mixes Kimi Delta Attention and MLA layers. It targets responsive agents, instruction following, and multi-turn chat at very low cost, with thinking on by default and switchable to an instant mode per request.
费用比率
0.1x
上下文
262K
发布日期
2026年8月6日
输入
Text
输出
Text
支持
推理工具调用结构化输出
最擅长
该模型排名最高的类别。
未找到排名靠前的类别。
性能
根据近期请求测得的中位延迟和吞吐量。
吞吐量183 tok/s
延迟2.17s