Unverified60% confidenceBenchmarkExact time
The arithmetic operations per token for Ling 3.0 Tiny are approximately one-sixth of those of an 8B dense model.
2
Sources
60%
Confidence
Medium-term (~90 days)
Relevance
8/10/2026
First Seen
Valid until: 11/8/2026
Sources
Related Claims
Unverified蚂蚁集团实验室发布了 Ling 3.0 Tiny 模型，总参数7.9B，每token激活约1.3B72% similarUnverified4-bit量化只能表示16个离散值，1-bit只有2个，三元量化（-1/0/+1）只有3个可能值64% similarUnverifiedOn the LibriSpeech benchmark, Oído's word error rate (WER) is 3.7 (clean) and 8.2 (other), while Whisper tiny.en on a laptop achieves 6.3 and 15.9.62% similarUnverified在4-bit量化下，纯参数理论最小占用为：7B约3.26GiB，13B约6.05GiB，32B约14.90GiB，70B约32.60GiB62% similarUnverifiedLing-3.0-flash adopts a Mixture of Experts (MoE) architecture with an activation ratio of only approximately 4% (approximately 5.1B activated out of 124B total parameters).62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/723989API
curl https://kongchang.com/api/v1/knowledge/claims/723989MCP
get_claim(id=723989)