Unverified50% confidenceBenchmarkExact time
In strict format scoring tests, TwIL-LM's 1.7B model reportedly outperforms models 5 to 15 times its size.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/11/2026
First Seen
Valid until: 12/10/2026
Sources
Related Entities
Related Claims
UnverifiedThe LLaMA series adopts an over-training strategy, training a 7B-parameter model on 1T tokens, far exceeding the Chinchilla-optimal ratio of 140B tokens67% similarUnverifiedLFM2.5在多项基准测试中展现出与体量高达其4倍模型相当的竞争力66% similarUnverifiedThe MiniCPM series achieves performance close to 7B and even 13B models on multiple benchmarks with a parameter count at the 2B–4B level65% similarUnverified研究表明当模型参数规模达到3B以上时，1.58-bit模型在多项基准测试中表现可接近甚至匹配同等规模的全精度模型65% similarUnverified万象模型提供了 1.3B 和 14B 参数量的不同规模版本64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/902361API
curl https://kongchang.com/api/v1/knowledge/claims/902361MCP
get_claim(id=902361)